Choroba zwyrodnieniowa stawów – Rokowania, prognozy i postęp choroby

Choroba zwyrodnieniowa stawów
Rokowania, prognozy i postęp choroby

Choroba zwyrodnieniowa stawów (osteoarthritis) stanowi istotne wyzwanie kliniczne, zwłaszcza u pacjentów powyżej 50. roku życia, gdzie precyzyjne prognozowanie progresji jest kluczowe dla optymalizacji terapii. Nowatorskie modele prognostyczne, oparte na zmiennych klinicznych i biomarkerach białkowych, umożliwiają przewidywanie radiograficznej choroby zwyrodnieniowej stawu kolanowego w okresie do 96 miesięcy, osiągając wartość AUC 0,83. Istotne czynniki prognostyczne obejmują obrzęk szpiku kostnego w MRI (iloraz szans 5,29, 95% CI 1,64-17,1, p=0,005), biomarkery takie jak CRTAC1 i COL9A1 oraz geny związane z sygnalizacją TGF-β. Czynniki kliniczne, takie jak wyższy wiek, BMI oraz stosowanie NLPZ, zwiększają ryzyko progresji, natomiast obecność przewlekłego bólu i objawów depresyjnych negatywnie wpływa na długoterminową odpowiedź na leczenie.

Prognoza choroby zwyrodnieniowej stawów

Modele prognostyczne w chorobie zwyrodnieniowej stawów
Czynniki prognostyczne w chorobie zwyrodnieniowej stawów
Prognozowanie odpowiedzi na leczenie
Zastosowanie uczenia maszynowego w prognozowaniu
Podgrupy pacjentów i heterogeniczność choroby

Wyzwania i ograniczenia w prognozowaniu
Perspektywy na przyszłość

Kolejne rozdziały

Prognoza choroby zwyrodnieniowej stawów

Choroba zwyrodnieniowa stawów (osteoarthritis) to jedna z najczęściej występujących chorób stawów, szczególnie u osób powyżej 50. roku życia. Precyzyjne przewidywanie progresji choroby i określenie rokowania staje się coraz ważniejszym elementem postępowania klinicznego. Umożliwia to wczesną interwencję, dobór odpowiednich strategii terapeutycznych oraz potencjalnie zmniejszenie obciążenia związanego z chorobą.¹²³

Modele prognostyczne w chorobie zwyrodnieniowej stawów

W ostatnich latach obserwuje się znaczący wzrost zainteresowania tworzeniem modeli prognostycznych dla choroby zwyrodnieniowej stawów. Zaawansowane modele wykorzystujące uczenie maszynowe, sztuczną inteligencję oraz analizę dużych zbiorów danych umożliwiają coraz dokładniejsze przewidywanie przebiegu choroby.⁴⁵

Wśród najważniejszych osiągnięć w tym zakresie należy wymienić:

Opracowanie i zewnętrzną walidację nowatorskiego modelu prognostycznego opartego na powszechnych zmiennych klinicznych i biomarkerach białkowych do przewidywania wystąpienia radiograficznej choroby zwyrodnieniowej stawu kolanowego w okresie 96 miesięcy u osób bez radiograficznych objawów choroby⁶⁷
Stworzenie nomogramu, który jest użytecznym narzędziem do stratyfikacji populacji wysokiego ryzyka, umożliwiającym personalizację strategii leczenia⁸
Model prognostyczny osiągający wartość pola pod krzywą ROC na poziomie 0,83, co wskazuje na dobrą zdolność różnicowania pacjentów z ryzykiem rozwoju choroby od osób zdrowych⁹

Czynniki prognostyczne w chorobie zwyrodnieniowej stawów

Identyfikacja kluczowych czynników prognostycznych ma fundamentalne znaczenie dla przewidywania przebiegu choroby. Badania wykazały, że różne biomarkery i parametry kliniczne mogą pomóc w prognozowaniu rozwoju i progresji choroby zwyrodnieniowej stawów.¹⁰¹¹

Biomarkery prognostyczne

Współczesne badania wskazują na istotną rolę biomarkerów w prognozowaniu przebiegu choroby zwyrodnieniowej stawów:

Obrzęk szpiku kostnego widoczny w badaniu MRI po urazie kolana jest silnym predyktorem nowego wystąpienia lub progresji zmian zwyrodnieniowych w stawie rzepkowo-udowym w badaniu kontrolnym po roku (iloraz szans 5,29, 95% CI 1,64-17,1, p=0,005)¹²
Kluczowe geny i szlaki związane z chorobą zwyrodnieniową stawów (np. GDF5 i sygnalizacja TGF-β) oraz specyficzne biomarkery (np. CRTAC1 i COL9A1) mają znaczenie prognostyczne¹³
Badania obrazowe MRI wykazują największy wpływ na identyfikację pacjentów z szybką progresją choroby zwyrodnieniowej stawów, co może być zastosowane do wczesnego prognozowania w praktyce klinicznej¹⁴

Czynniki kliniczne i demograficzne

Wśród najważniejszych czynników klinicznych i demograficznych wpływających na prognozę wyróżnia się:

Wyższy wiek, większy BMI oraz przyjmowanie niesteroidowych leków przeciwzapalnych – te czynniki w największym stopniu przyczyniają się do zwiększonego ryzyka rozwoju choroby zwyrodnieniowej stawów¹⁵
Gorsza wyjściowa funkcja stawu i bardziej zaawansowane radiologiczne zmiany zwyrodnieniowe wiążą się z większą poprawą po leczeniu, ale pacjenci ci nigdy nie osiągają takiego poziomu funkcjonowania jak osoby z lepszą wyjściową funkcją lub mniej nasilonymi zmianami radiologicznymi¹⁶
Tkliwość wokół stawu kolanowego wiąże się z lepszym krótkoterminowym wynikiem po dostawowych iniekcjach steroidów, jednak czynniki kliniczne nie pozwalają przewidzieć odpowiedzi długoterminowej¹⁷
Obecność przewlekłego uogólnionego bólu, bólu w wielu miejscach oraz objawów depresyjnych osłabia długoterminową odpowiedź na leczenie¹⁸

Prognozowanie odpowiedzi na leczenie

Zdolność przewidywania odpowiedzi na leczenie ma kluczowe znaczenie dla optymalizacji interwencji terapeutycznych u pacjentów z chorobą zwyrodnieniową stawów.¹⁹²⁰

Badania wykazują, że:

Kliniczny nomogram składający się z czterech prostych pomiarów (niższe oczekiwane korzyści zgłaszane przez pacjenta, niższa zgłaszana przez pacjenta funkcja kolana, większy kąt szpotawości kolana i ciężka przyśrodkowa radiologiczna degeneracja kolana) może pomóc w identyfikacji pacjentów zagrożonych słabą odpowiedzią na niechirurgiczne leczenie wielodyscyplinarne²¹²²
Przedoperacyjny poziom bólu słabo koreluje z bólem resztkowym po zabiegu całkowitej wymiany stawu, co sugeruje, że pooperacyjny ból jest minimalnie związany z przedoperacyjnymi właściwościami bólu²³²⁴
Selekcja pacjentów oparta na modelach przewidujących progresję choroby może zmniejszyć o 20-25% liczbę pacjentów niewykazujących progresji w badaniach klinicznych, co może poprawić ich efektywność²⁵²⁶

Zastosowanie uczenia maszynowego w prognozowaniu

Wykorzystanie zaawansowanych technik uczenia maszynowego i sztucznej inteligencji otwiera nowe możliwości w prognozowaniu przebiegu choroby zwyrodnieniowej stawów.²⁷²⁸

Kluczowe osiągnięcia w tym zakresie obejmują:

Model przewidujący wystąpienie całkowitej wymiany stawu kolanowego w ciągu 2 i 5 lat u pacjentów z chorobą zwyrodnieniową stawów, wykorzystujący rutynowo zbierane dane pacjentów, osiągnął akceptowalny klinicznie poziom dokładności (AUC>0,7)²⁹
Kombinacja danych z obu kończyn dolnych w modelach uczenia maszynowego pozwala na osiągnięcie najlepszej wydajności – model regresji logistycznej osiągnął dokładność 83,3% przy znacznie mniejszej liczbie cech (29)³⁰
Analiza wykazała, że mieszanka heterogenicznych cech z niemal wszystkich kategorii jest niezbędna w celu maksymalizacji wydajności i dokładności prognozowania modeli³¹

W projekcie KNOAP2020 (KNee OsteoArthritis Prediction) opracowano modele wykorzystujące konwolucyjne sieci neuronowe (CNN) do ekstrakcji informacji z obrazów rentgenowskich i połączenia tych informacji ze zmiennymi klinicznymi (wiek, BMI, stopień KL), osiągając najwyższy wskaźnik ROC AUC (0,64).³²

Podgrupy pacjentów i heterogeniczność choroby

Choroba zwyrodnieniowa stawów charakteryzuje się dużą heterogenicznością, co wpływa na jej prognozę. Identyfikacja podgrup pacjentów o różnych profilach ryzyka może przyczynić się do lepszego prognozowania i personalizacji leczenia.³³³⁴

Badania wskazują, że:

Zidentyfikowano 14 podgrup profili ryzyka choroby zwyrodnieniowej stawów, które zostały zwalidowane w niezależnym zbiorze pacjentów oceniających 11-letnie ryzyko, przy czym 88% pacjentów zostało jednoznacznie przypisanych do jednej z 14 podgrup³⁵
Indywidualne profile ryzyka choroby zwyrodnieniowej stawów charakteryzują się spersonalizowanymi biomarkerami³⁶
Czynniki biologiczne i środowiskowe leżące u podstaw choroby zwyrodnieniowej stawów są heterogenne wśród poszczególnych osób, co zostało wykazane przez grupowanie wartości SHAP oszacowanych przez model kliniczny³⁷

Wyzwania i ograniczenia w prognozowaniu

Pomimo znaczących postępów w dziedzinie prognozowania choroby zwyrodnieniowej stawów, istnieją pewne wyzwania i ograniczenia, które należy wziąć pod uwagę.³⁸³⁹

Dokładne przewidywanie objawowej radiograficznej choroby zwyrodnieniowej stawu kolanowego jest złożonym i nadal nierozwiązanym problemem wymagającym dodatkowych badań⁴⁰
Modele opublikowane do tej pory koncentrują się głównie na chorobie zwyrodnieniowej stawu kolanowego i opierają się na stosunkowo małej liczbie podstawowych zbiorów danych kohortowych⁴¹
Wiele modeli nie jest zaprojektowanych (ani jeszcze nie przeznaczonych) do masowego zastosowania⁴²
Literatura dotycząca binarnej klasyfikacji objawowej choroby zwyrodnieniowej stawu kolanowego zapewnia wysoką dokładność, jednak dalsza walidacja w celu zminimalizowania przeuczenia jest wymagana⁴³
Obszary prognozowania i wieloklasowej klasyfikacji stopnia zaawansowania choroby zwyrodnieniowej stawów pozostają obszarami do dalszego rozwoju⁴⁴

Perspektywy na przyszłość

Przewidywanie przebiegu choroby zwyrodnieniowej stawów jest dynamicznie rozwijającą się dziedziną, która stwarza nowe możliwości w zakresie wczesnej interwencji i personalizacji leczenia.⁴⁵⁴⁶

Przyszłe kierunki badań w tej dziedzinie obejmują:

Opracowywanie spersonalizowanych strategii zapobiegawczych opartych na indywidualnych profilach ryzyka⁴⁷
Wykorzystanie modeli prognostycznych do poprawy rekrutacji pacjentów do badań klinicznych⁴⁸
Dłuższy okres obserwacji uczestników w obecnych badaniach, w tym zarówno wyników klinicznych, jak i chirurgicznych⁴⁹
Wyprowadzenie markerów prognostycznych dla utrzymywania się bólu po całkowitej wymianie stawu, co będzie wymagać bardziej kompleksowego zrozumienia podstawowych mechanizmów⁵⁰
Rozwój modeli uwzględniających heterogeniczność choroby zwyrodnieniowej stawów i identyfikujących podgrupy pacjentów o różnych profilach ryzyka⁵¹

Wykorzystanie zaawansowanych technik analizy danych w połączeniu z licznymi biomarkerami i danymi klinicznymi otwiera nowe możliwości w prognozowaniu przebiegu choroby zwyrodnieniowej stawów, umożliwiając bardziej precyzyjne przewidywanie progresji choroby i odpowiedzi na leczenie, co w konsekwencji może prowadzić do poprawy jakości życia pacjentów.⁵²⁵³

Kolejne rozdziały

Zapraszamy do dalszego czytania naszego leksykonu.

Wybierz kolejny rozdział z menu poniżej, aby otworzyć nową podstronę kompedium wiedzy i uzyskać szczegółowe informację o leku, substancji lub chorobie.

Materiały źródłowe

#1 Prognostic model to predict the incidence of radiographic knee osteoarthritis | Annals of the Rheumatic Diseases
https://ard.bmj.com/content/83/5/661
Osteoarthritis Prognostic model to predict the incidence of radiographic knee osteoarthritis […] Early diagnosis of knee osteoarthritis (KOA) in asymptomatic stages is essential for the timely management of patients using preventative strategies. We develop and validate a prognostic model useful for predicting the incidence of radiographic KOA (rKOA) in non-radiographic osteoarthritic subjects and stratify individuals at high risk of developing the disease. […] A novel prognostic model based on common clinical variables and protein biomarkers was developed and externally validated to predict rKOA incidence over a 96-month period in individuals without any radiographic signs of disease. The resulting nomogram is a useful tool for stratifying high-risk populations and could potentially lead to personalised medicine strategies for treating OA.
#2 Prognostic model to predict the incidence of radiographic knee osteoarthritis – PubMed
https://pubmed.ncbi.nlm.nih.gov/38182405/
Objective: Early diagnosis of knee osteoarthritis (KOA) in asymptomatic stages is essential for the timely management of patients using preventative strategies. We develop and validate a prognostic model useful for predicting the incidence of radiographic KOA (rKOA) in non-radiographic osteoarthritic subjects and stratify individuals at high risk of developing the disease. […] Conclusion: A novel prognostic model based on common clinical variables and protein biomarkers was developed and externally validated to predict rKOA incidence over a 96-month period in individuals without any radiographic signs of disease. The resulting nomogram is a useful tool for stratifying high-risk populations and could potentially lead to personalised medicine strategies for treating OA.
#3
https://urncst.com/index.php/urncst/article/view/564
Knee Osteoarthritis (KOA) is the second most reported condition for persons 50 years and up; approximated by the continuous degradation of the knee, and eventually extending to the debilitation of biomechanical gait parameters. […] Of the 22 articles, 10% focused on the prediction of patient outcomes and disease prognosis, while 90% focused on the initial diagnosis or severity prediction of KOA. […] Overall, literature concerning the binary classification of symptomatic KOA provided high accuracy, yet further validation to minimize overfitting is required. Furthermore, areas for prognosis prediction and multiclass classification of KOA severity remain as areas for further development.
#4 PREDICTION MODELS TO ESTIMATE FUTURE INDIVIDUAL RISK OF OSTEOARTHRITIS IN THE GENERAL POPULATION: A SYSTEMATIC REVIEW
https://keele-repository.worktribe.com/output/507164/prediction-models-to-estimate-future-individual-risk-of-osteoarthritis-in-the-general-population-a-systematic-review
Purpose: The need for health systems to shift from late, reactive care of osteoarthritis (OA) to earlier, preventative strategies is widely acknowledged. […] Our objective was to critically synthesise published evidence on the performance of multivariable prediction models for OA incidence and their applicability to large-scale use in the general population. […] The most common outcome was incident OA defined by plain radiography. […] The median prediction horizon was 8 years (range 2 to 41 years), median number of participants/joints with the outcome of interest was 99 (range 27 to 12,803), and the median number of predictors included in the final models was 5.5 (range 3 to 13). […] Model performance for 25 of the 26 models was presented by Area under the Curve (AUC). Median performance for knee, hip and hand OA was 0.72, 0.76, 0.62, respectively. The one model for any-site OA had an AUC of 0.84.
#5 PREDICTION MODELS TO ESTIMATE FUTURE INDIVIDUAL RISK OF OSTEOARTHRITIS IN THE GENERAL POPULATION: A SYSTEMATIC REVIEW
https://keele-repository.worktribe.com/output/507164/prediction-models-to-estimate-future-individual-risk-of-osteoarthritis-in-the-general-population-a-systematic-review
Conclusions: Of the 21 studies found and included in our review, 15 were published within the past 4 years, suggesting increasing interest among researchers in predicting individual-level risk for OA. […] However, models published to date remain heavily focussed on knee OA and have relied on a relatively small number of underlying cohort datasets. […] Furthermore, our systematic review highlights common shortfalls in applicability, suggesting that many models are not designed (nor yet intended) for mass application.
#6 Prognostic model to predict the incidence of radiographic knee osteoarthritis | Annals of the Rheumatic Diseases
https://ard.bmj.com/content/83/5/661
Osteoarthritis Prognostic model to predict the incidence of radiographic knee osteoarthritis […] Early diagnosis of knee osteoarthritis (KOA) in asymptomatic stages is essential for the timely management of patients using preventative strategies. We develop and validate a prognostic model useful for predicting the incidence of radiographic KOA (rKOA) in non-radiographic osteoarthritic subjects and stratify individuals at high risk of developing the disease. […] A novel prognostic model based on common clinical variables and protein biomarkers was developed and externally validated to predict rKOA incidence over a 96-month period in individuals without any radiographic signs of disease. The resulting nomogram is a useful tool for stratifying high-risk populations and could potentially lead to personalised medicine strategies for treating OA.
#7 Prognostic model to predict the incidence of radiographic knee osteoarthritis | Annals of the Rheumatic Diseases
https://ard.bmj.com/content/83/5/661
A prognostic model based on clinical data and three protein biomarkers was developed and externally validated to predict the incidence of radiographic KOA in individuals without any radiographic signs of the disease, with an area under the curve of 0.83. […] This nomogram for KOA incidence is a useful tool for stratifying high-risk populations in order to prevent disease onset or delay its progression, and thereby decrease the associated burden. The prognostic model may also be valuable in improving patient recruitment into clinical trials.
#8 Prognostic model to predict the incidence of radiographic knee osteoarthritis | Annals of the Rheumatic Diseases
https://ard.bmj.com/content/83/5/661
A prognostic model based on clinical data and three protein biomarkers was developed and externally validated to predict the incidence of radiographic KOA in individuals without any radiographic signs of the disease, with an area under the curve of 0.83. […] This nomogram for KOA incidence is a useful tool for stratifying high-risk populations in order to prevent disease onset or delay its progression, and thereby decrease the associated burden. The prognostic model may also be valuable in improving patient recruitment into clinical trials.
#9 Prognostic model to predict the incidence of radiographic knee osteoarthritis | Annals of the Rheumatic Diseases
https://ard.bmj.com/content/83/5/661
A prognostic model based on clinical data and three protein biomarkers was developed and externally validated to predict the incidence of radiographic KOA in individuals without any radiographic signs of the disease, with an area under the curve of 0.83. […] This nomogram for KOA incidence is a useful tool for stratifying high-risk populations in order to prevent disease onset or delay its progression, and thereby decrease the associated burden. The prognostic model may also be valuable in improving patient recruitment into clinical trials.
#10 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
Osteoarthritis (OA) is increasing in prevalence and has a severe impact on patients lives. However, our understanding of biomarkers driving OA risk remains limited. We developed a model predicting the five-year risk of OA diagnosis, integrating retrospective clinical, lifestyle and biomarker data from the UK Biobank (19,120 patients with OA, ROC-AUC: 0.72, 95%CI (0.710.73)). Higher age, BMI and prescription of non-steroidal anti-inflammatory drugs contributed most to increased OA risk prediction ahead of diagnosis. We identified 14 subgroups of OA risk profiles. These subgroups were validated in an independent set of patients evaluating the 11-year OA risk, with 88% of patients being uniquely assigned to one of the 14 subgroups. Individual OA risk profiles were characterised by personalised biomarkers. Omics integration demonstrated the predictive importance of key OA genes and pathways (e.g., GDF5 and TGF- signalling) and OA-specific biomarkers (e.g., CRTAC1 and COL9A1). In summary, this work identifies opportunities for personalised OA prevention and insights into its underlying pathogenesis.
#11 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
In this retrospective study, we develop a machine learning model to predict individual risk and identify risk biomarkers up to 5-years prior to an OA diagnosis. Through the integration of multi-modal patient data, we identify subgroups of OA, with different risk biomarker profiles, which is validated to be effective on an unseen subpopulation of the UK Biobank up to 11 years ahead of diagnosis. The model captures the broad risk biomarker landscape, in a UK cohort of ~20,000 people diagnosed with OA, utilising electronic health records (EHR), clinical biomarkers, self-reported questionnaire data, genomics, proteomics, and metabolomics on available subsets of individuals. The model quantifies the impact of risk biomarkers on the predicted OA risk at the population and individual level, enabling detailed estimation of the contribution of these biomarkers for OA risk.
#12
https://link.springer.com/article/10.1007/s00330-011-2089-3
To prospectively evaluate prognostic factors for new onset or progression of degenerative change on follow-up MRI one year after knee trauma and the association with clinical outcome. […] The only statistically significant prognostic variable in the multivariate analysis was bone marrow oedema on initial MRI (OR 5.29 (95% CI 1.6417.1), p=0.005). […] Bone marrow oedema on MRI for acute knee injury is strongly predictive of new onset or progression of degenerative change of the femorotibial joint on follow-up MRI one year after trauma, which is reflected in clinical outcome. […] In conclusion, the results of this study demonstrate that bone marrow oedema on initial MRI after knee trauma is a strong predictor, and in multivariable analysis the only predictor, of new onset and progression of knee osteoarthritis on 1-year follow-up MRI, which is reflected in the clinical outcome.
#13 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
Osteoarthritis (OA) is increasing in prevalence and has a severe impact on patients lives. However, our understanding of biomarkers driving OA risk remains limited. We developed a model predicting the five-year risk of OA diagnosis, integrating retrospective clinical, lifestyle and biomarker data from the UK Biobank (19,120 patients with OA, ROC-AUC: 0.72, 95%CI (0.710.73)). Higher age, BMI and prescription of non-steroidal anti-inflammatory drugs contributed most to increased OA risk prediction ahead of diagnosis. We identified 14 subgroups of OA risk profiles. These subgroups were validated in an independent set of patients evaluating the 11-year OA risk, with 88% of patients being uniquely assigned to one of the 14 subgroups. Individual OA risk profiles were characterised by personalised biomarkers. Omics integration demonstrated the predictive importance of key OA genes and pathways (e.g., GDF5 and TGF- signalling) and OA-specific biomarkers (e.g., CRTAC1 and COL9A1). In summary, this work identifies opportunities for personalised OA prevention and insights into its underlying pathogenesis.
#14 Developing a Comprehensive Patient-Specific Disease Progression Prediction Model for Knee Osteoarthritis Using Machine/Deep Learning Methods – ACR Meeting Abstracts
https://acrabstracts.org/abstract/developing-a-comprehensive-patient-specific-disease-progression-prediction-model-for-knee-osteoarthritis-using-machine-deep-learning-methods/
Developing a Comprehensive Patient-Specific Disease Progression Prediction Model for Knee Osteoarthritis Using Machine/Deep Learning Methods […] Current assessment of knee osteoarthritis (OA) is primarily based on a patients personal and familial history, clinical features, and on radiographies. However, such information does not provide enough evidence to lead to a robust prediction/prognosis of OA fast progressors. This study aims to identify early significant predictors of knee OA progression using advanced machine learning (ML) and deep learning (DL) algorithms and propose a prediction model based on a short number of selected predictors and outcomes. […] This is the first time that such a comprehensive study is performed for identifying the best predictors of knee OA rapid progressors. Importantly, data showed that MRI-based variables and outcome have the most significant impact in identifying OA progressors, and could be applied for early prognosis in clinical practice.
#15 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
Osteoarthritis (OA) is increasing in prevalence and has a severe impact on patients lives. However, our understanding of biomarkers driving OA risk remains limited. We developed a model predicting the five-year risk of OA diagnosis, integrating retrospective clinical, lifestyle and biomarker data from the UK Biobank (19,120 patients with OA, ROC-AUC: 0.72, 95%CI (0.710.73)). Higher age, BMI and prescription of non-steroidal anti-inflammatory drugs contributed most to increased OA risk prediction ahead of diagnosis. We identified 14 subgroups of OA risk profiles. These subgroups were validated in an independent set of patients evaluating the 11-year OA risk, with 88% of patients being uniquely assigned to one of the 14 subgroups. Individual OA risk profiles were characterised by personalised biomarkers. Omics integration demonstrated the predictive importance of key OA genes and pathways (e.g., GDF5 and TGF- signalling) and OA-specific biomarkers (e.g., CRTAC1 and COL9A1). In summary, this work identifies opportunities for personalised OA prevention and insights into its underlying pathogenesis.
#16 Preoperative predictors for outcomes after total hip replacement in patients with osteoarthritis: a systematic review | BMC Musculoskeletal Disorders | Full Text
https://bmcmusculoskeletdisord.biomedcentral.com/articles/10.1186/s12891-016-1070-3
This systematic review examines which patient related factors influence functional and clinical outcomes after total hip arthroplasty (THA) in patients with hip osteoarthritis (OA). […] Overall, preoperative function (13 studies (37 %), 2 with low risk of bias) and radiological OA (6 studies (17 %), 1 with low risk of bias) were predictors with the most consistent findings. Worse preoperative function and more severe radiological OA were associated with larger postoperative improvement. However, these patients never reached the level of postoperative functioning as patients with better preoperative function or less severe radiological OA. […] There is not enough evidence to draw succinct conclusions on preoperative predictors for postoperative outcome in THA, as results of studies are conflicting and the methodological quality is low. Results suggest to focus on preoperative function and radiological osteoarthritis to decide when THA will be most effective. […] Overall, even though greater improvements were found in patients with more severe radiological OA and lower function baseline scores, these patients did not reach the same postoperative levels in functioning as patients with less severe OA or higher baseline function scores.
#17 Do Clinical Correlates of Knee Osteoarthritis Predict Outcome of Intraarticular Steroid Injections? | The Journal of Rheumatology
https://www.jrheum.org/content/47/3/431
Objective. To determine whether clinical correlates of knee osteoarthritis (OA) affect the outcome of intraarticular steroid injections (IASI) in symptomatic knee OA. […] Conclusion. Among patients with symptomatic knee OA, tenderness around the knee was associated with better short-term outcome of IASI. However, clinical-related factors did not predict longer-term response, while those with chronic widespread pain and depressive symptoms were less likely to obtain longer-term benefits. […] Among patients with symptomatic knee OA, those with knee tenderness are more likely to respond to IASI therapy. Clinical signs of knee OA did not, however, predict longer-term response. The presence of CWP, having multiple pain sites, and depressive symptoms attenuate longer-term treatment response.
#18 Do Clinical Correlates of Knee Osteoarthritis Predict Outcome of Intraarticular Steroid Injections? | The Journal of Rheumatology
https://www.jrheum.org/content/47/3/431
Objective. To determine whether clinical correlates of knee osteoarthritis (OA) affect the outcome of intraarticular steroid injections (IASI) in symptomatic knee OA. […] Conclusion. Among patients with symptomatic knee OA, tenderness around the knee was associated with better short-term outcome of IASI. However, clinical-related factors did not predict longer-term response, while those with chronic widespread pain and depressive symptoms were less likely to obtain longer-term benefits. […] Among patients with symptomatic knee OA, those with knee tenderness are more likely to respond to IASI therapy. Clinical signs of knee OA did not, however, predict longer-term response. The presence of CWP, having multiple pain sites, and depressive symptoms attenuate longer-term treatment response.
#19 Prospective validity of a clinical prediction rule for response to non-surgical multidisciplinary management of knee osteoarthritis in tertiary care: a multisite prospective longitudinal study | BMJ Open
https://bmjopen.bmj.com/content/14/3/e078531
We tested a previously developed clinical prediction toola nomogram consisting of four patient measures (lower patient-expected benefit, lower patient-reported knee function, greater knee varus angle and severe medial knee radiological degeneration) that were related to poor response to non-surgical management of knee osteoarthritis. This study sought to prospectively evaluate the predictive validity of this nomogram to identify patients most likely to respond poorly to non-surgical management of knee osteoarthritis. […] The knee osteoarthritis clinical nomogram prediction tool may have capacity to identify patients at risk of poor response to non-surgical management. Further work is required to determine the implications for service delivery, feasibility and impact of implementing the nomogram in clinical practice.
#20 Prospective validity of a clinical prediction rule for response to non-surgical multidisciplinary management of knee osteoarthritis in tertiary care: a multisite prospective longitudinal study | BMJ Open
https://bmjopen.bmj.com/content/14/3/e078531
The capacity to identify individuals at risk of a poor response to non-surgical multidisciplinary care of their knee osteoarthritis would be advantageous. Potentially individuals identified to be at risk of a poor response may benefit from a matched or stratified care approach where they may be provided more tailored care to address identified risk factors of poor response. […] This study evaluated the prospective validity of a previously developed clinical prediction rule in the form of a nomogram to identify patients at risk of poor response to non-surgical multidisciplinary management of knee osteoarthritis. At optimal combined specificity/sensitivity, the clinical nomogram demonstrated low capacity to predict response to non-surgical management, with modest positive and negative LRs.
#21 Prospective validity of a clinical prediction rule for response to non-surgical multidisciplinary management of knee osteoarthritis in tertiary care: a multisite prospective longitudinal study | BMJ Open
https://bmjopen.bmj.com/content/14/3/e078531
We tested a previously developed clinical prediction toola nomogram consisting of four patient measures (lower patient-expected benefit, lower patient-reported knee function, greater knee varus angle and severe medial knee radiological degeneration) that were related to poor response to non-surgical management of knee osteoarthritis. This study sought to prospectively evaluate the predictive validity of this nomogram to identify patients most likely to respond poorly to non-surgical management of knee osteoarthritis. […] The knee osteoarthritis clinical nomogram prediction tool may have capacity to identify patients at risk of poor response to non-surgical management. Further work is required to determine the implications for service delivery, feasibility and impact of implementing the nomogram in clinical practice.
#22 Prospective validity of a clinical prediction rule for response to non-surgical multidisciplinary management of knee osteoarthritis in tertiary care: a multisite prospective longitudinal study | BMJ Open
https://bmjopen.bmj.com/content/14/3/e078531
A clinical nomogram consisting of four simple measures to identify patients unlikely to respond to multidisciplinary non-surgical care for knee osteoarthritis in tertiary care has demonstrated capacity to predict poor response to non-surgical care, particularly when using higher cut-off scores. Further work is required to include a longer-term follow-up of participants in the current study, including both clinical and surgical outcomes.
#23 Prognostics for pain in osteoarthritis: Do clinical measures predict pain after total joint replacement? | PLOS One
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0222370
A significant proportion of osteoarthritis (OA) patients continue to experience moderate to severe pain after total joint replacement (TJR). […] Pre-surgical pain levels were not related to post-surgical residual pain. […] Our results show that although tested clinical and biopsychosocial variables reorganize after TJR in OA, their presurgical values are not predictive of post-surgery pain. […] Derivation of prognostic markers for pain persistence after TJR will require more comprehensive understanding of underlying mechanisms. […] Pain intensity prior to surgery, disproportion between pain intensity and articular damage, neuropathic-like symptoms, psychosocial factors such as pain catastrophizing and poor coping strategies are commonly referenced as important predictive factors. […] Our main aims are: 1) Test if baseline pain ratings relate to post-surgery pain levels; 2) examine how distinct pain measurement instruments relate to different clinical and biopsychological aspects of OA pain; 3) develop and evaluate models predictive of pain and pain relief after TJR; 4) use network analysis to assess the reorganization of pain related clinical and biopsychosocial properties of the personality of KOA and HOA patients after TJR.
#24 Prognostics for pain in osteoarthritis: Do clinical measures predict pain after total joint replacement? | PLOS One
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0222370
For all 4 pain outcome measures, we examined the relationship between pre-surgical pain and a) residual pain after surgery (100% residual pain meaning no change; 0% residual pain rendering complete relief); b) absolute pain values after surgery, both for KOA and HOA at 3- and 6-months. […] We observed mostly weak and statistically not significant correlations between pre-surgical pain intensity and residual pain. […] Thus, our analysis, especially for KOA where we examined multiple models, suggests that the post-operative pain is minimally related to the pre-operative pain properties.
#25 Multi-classifier prediction of knee osteoarthritis progression from incomplete imbalanced longitudinal data | Scientific Reports
https://www.nature.com/articles/s41598-020-64643-8
Conventional inclusion criteria used in osteoarthritis clinical trials are not very effective in selecting patients who would benefit from a therapy being tested. […] This could be avoided, if selection criteria were more predictive of the future disease progression. […] We found that the model-based selection outperforms the conventional inclusion criteria, reducing by 20-25% the number of patients who show no progression. This result might lead to more efficient clinical trials. […] The main hypothesis of this article is that machine learning can be more effective at identifying progressive patients than the traditional approach. […] We hypothesise that prediction models trained on historical data will be able to differentiate between patients for whom a fast progression happen during the observation period, and patients who show no progression or progress slowly and should not be selected to trials.
#26 Multi-classifier prediction of knee osteoarthritis progression from incomplete imbalanced longitudinal data | Scientific Reports
https://www.nature.com/articles/s41598-020-64643-8
We observed a substantial reduction in the number of undesired non-progressive cases. This findings could impact the future clinical trials design, and potentially improve their efficiency. […] The model-based selection, compared to the conventional criteria, resulted in 20-25% less non-progressive cases and more balanced retrieval of progressive cases.
#27 Prediction of Joint Space Narrowing Progression in Knee Osteoarthritis Patients
https://www.mdpi.com/2075-4418/11/2/285
Osteoarthritis is a joint disease that commonly occurs in the knee (KOA). The continuous increase in medical data regarding KOA has triggered researchers to incorporate artificial intelligence analytics for KOA prognosis or treatment. In this study, two approaches are presented to predict the progression of knee joint space narrowing (JSN) in each knee and in both knees combined. The proposed methodology employs: (i) A clustering process to identify groups of people with progressing and non-progressing JSN; (ii) a robust feature selection (FS) process consisting of filter, wrapper, and embedded techniques that identifies the most informative risk factors; (iii) a decision making process based on the evaluation and comparison of various classification algorithms towards the selection and development of the final predictive model for JSN; and (iv) post-hoc interpretation of the featuresâ impact on the best performing model. The results showed that bounding the JSN progression of both knees can result to more robust prediction models with a higher accuracy (83.3%) and with fewer risk factors (29) compared to the right knee (77.7%, 88 risk factors) and the left knee (78.3%, 164 risk factors), separately.
#28 Prediction of Joint Space Narrowing Progression in Knee Osteoarthritis Patients
https://www.mdpi.com/2075-4418/11/2/285
The diagnosis or even the treatment of KOA is still a challenge for the scientific community. However, the increasing amount of medical data related to KOA permitted the development of more recent studies by using artificial intelligence and big data. […] This study constitutes the first attempt to a large-scale integration of skeletal biomechanics and compositional imaging. […] The proposed method achieved an 82.98% accuracy. […] The main objective of this study was the accurate prediction of JSN in KOA patients based on a machine learning pipeline trained on multimodal data from the OAI (725 features in total were considered). […] The outcome of the ML models indicated that the LR model achieved the best performance for the left leg with a 78.3% accuracy for 164 features, while for the right leg, the SVM model dominated with a 77.7% accuracy for 88 and 90 features.
#29 Predicting total knee replacement at 2 and 5 years in osteoarthritis patients using machine learning – LSE Research Online
http://eprints.lse.ac.uk/124424/
Objectives Knee osteoarthritis is a major cause of physical disability and reduced quality of life, with end-stage disease often treated by total knee replacement (TKR). […] The primary outcome of this study was prediction of TKR onset at 2 and 5 years. […] Our approach suggests that routinely collected patient data are sufficient to drive a predictive model with a clinically acceptable level of accuracy (AUC0.7) and is the first such tool to be externally validated. This level of accuracy is higher than previously published models utilising MRI data, which is not routinely collected.
#30 Prediction of Joint Space Narrowing Progression in Knee Osteoarthritis Patients
https://www.mdpi.com/2075-4418/11/2/285
However, the best overall performance was achieved by the second strategy where the data from both legs were combined. Specifically, the LR model achieved a 83.3% accuracy for a significantly lower number of features (29). […] Through this analysis, we concluded that a blend of heterogeneous features from almost all feature categories is necessary in order to maximize the performance and prediction accuracy of the models.
#31 Prediction of Joint Space Narrowing Progression in Knee Osteoarthritis Patients
https://www.mdpi.com/2075-4418/11/2/285
However, the best overall performance was achieved by the second strategy where the data from both legs were combined. Specifically, the LR model achieved a 83.3% accuracy for a significantly lower number of features (29). […] Through this analysis, we concluded that a blend of heterogeneous features from almost all feature categories is necessary in order to maximize the performance and prediction accuracy of the models.
#32 The KNee OsteoArthritis Prediction (KNOAP2020) challenge: An image analysis challenge to predict incident symptomatic radiographic knee osteoarthritis from MRI and X-ray images
https://pmc.ncbi.nlm.nih.gov/articles/PMC10323696/
The KNee OsteoArthritis Prediction (KNOAP2020) challenge was organized to objectively compare methods for the prediction of incident symptomatic radiographic knee osteoarthritis within 78 months on a test set with blinded ground truth. […] The KNOAP2020 challenge established a benchmark for predicting incident symptomatic radiographic knee osteoarthritis. Accurate prediction of incident symptomatic radiographic knee osteoarthritis is a complex and still unsolved problem requiring additional investigation. […] The aim of this challenge was to objectively compare different methods for the prediction of incident symptomatic radiographic knee OA (according to the combined American College of Rheumatology (ACR) criteria) within 78 months on a test set with blinded ground truth. […] The model with the highest ROC AUC (0.64) used a CNN-based model to extract information from X-ray images and combined that information with clinical variables (i.e., age, BMI, and KL grade). The model with the highest BACC (0.59) ensembled three different models that used automatically extracted X-ray and MRI features along with clinical variables. […] The performance of the submitted models on the independent test set with blinded ground truth was limited indicating that accurate prediction of incident symptomatic radiographic knee OA is a complex and still unsolved problem that requires additional investigation.
#33 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
In this retrospective study, we develop a machine learning model to predict individual risk and identify risk biomarkers up to 5-years prior to an OA diagnosis. Through the integration of multi-modal patient data, we identify subgroups of OA, with different risk biomarker profiles, which is validated to be effective on an unseen subpopulation of the UK Biobank up to 11 years ahead of diagnosis. The model captures the broad risk biomarker landscape, in a UK cohort of ~20,000 people diagnosed with OA, utilising electronic health records (EHR), clinical biomarkers, self-reported questionnaire data, genomics, proteomics, and metabolomics on available subsets of individuals. The model quantifies the impact of risk biomarkers on the predicted OA risk at the population and individual level, enabling detailed estimation of the contribution of these biomarkers for OA risk.
#34 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
The Clin model confirmed that the biological and environmental risk factors underlying OA are heterogenous across individuals. We attempted to capture this heterogeneity and categorise patients into subgroups with differing risk biomarker profiles. Hence, we clustered the SHAP values, as estimated by the Clin model, for all risk biomarkers across all individuals. The clustering allowed us to uncover subgroups of individuals predicted to have high risk of OA. […] We present a large-scale study of OA, in a UK cohort of ~20,000 patients with OA and ~20,000 controls, encompassing a broad set of longitudinal OA risk biomarkers. We utilised interpretable machine learning to address gaps in the field of prediction and understanding of OA pathogenesis. Our models advance our understanding of risk of OA diagnosis, generating hypotheses for early screening, prevention and treatment of OA. The presented models predict an individuals 5-year risk of an OA diagnosis from EHR data which encompasses the range of OA heterogeneity and pathophysiology in real-life clinical settings. The complexity of OA risk was captured using multi-modal clinical data, biochemical and molecular signatures of OA. We identified distinct subgroups of OA risk profiles and derived simple clinical association rules for these subgroups. We also mapped differentially expressed molecular biomarkers between OA cases in risk subgroups. Finally, we demonstrated how individual patient journeys can be dissected to identify risk biomarkers for OA, contributing to the development of personalised preventative strategies.
#35 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
Osteoarthritis (OA) is increasing in prevalence and has a severe impact on patients lives. However, our understanding of biomarkers driving OA risk remains limited. We developed a model predicting the five-year risk of OA diagnosis, integrating retrospective clinical, lifestyle and biomarker data from the UK Biobank (19,120 patients with OA, ROC-AUC: 0.72, 95%CI (0.710.73)). Higher age, BMI and prescription of non-steroidal anti-inflammatory drugs contributed most to increased OA risk prediction ahead of diagnosis. We identified 14 subgroups of OA risk profiles. These subgroups were validated in an independent set of patients evaluating the 11-year OA risk, with 88% of patients being uniquely assigned to one of the 14 subgroups. Individual OA risk profiles were characterised by personalised biomarkers. Omics integration demonstrated the predictive importance of key OA genes and pathways (e.g., GDF5 and TGF- signalling) and OA-specific biomarkers (e.g., CRTAC1 and COL9A1). In summary, this work identifies opportunities for personalised OA prevention and insights into its underlying pathogenesis.
#36 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
Osteoarthritis (OA) is increasing in prevalence and has a severe impact on patients lives. However, our understanding of biomarkers driving OA risk remains limited. We developed a model predicting the five-year risk of OA diagnosis, integrating retrospective clinical, lifestyle and biomarker data from the UK Biobank (19,120 patients with OA, ROC-AUC: 0.72, 95%CI (0.710.73)). Higher age, BMI and prescription of non-steroidal anti-inflammatory drugs contributed most to increased OA risk prediction ahead of diagnosis. We identified 14 subgroups of OA risk profiles. These subgroups were validated in an independent set of patients evaluating the 11-year OA risk, with 88% of patients being uniquely assigned to one of the 14 subgroups. Individual OA risk profiles were characterised by personalised biomarkers. Omics integration demonstrated the predictive importance of key OA genes and pathways (e.g., GDF5 and TGF- signalling) and OA-specific biomarkers (e.g., CRTAC1 and COL9A1). In summary, this work identifies opportunities for personalised OA prevention and insights into its underlying pathogenesis.
#37 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
The Clin model confirmed that the biological and environmental risk factors underlying OA are heterogenous across individuals. We attempted to capture this heterogeneity and categorise patients into subgroups with differing risk biomarker profiles. Hence, we clustered the SHAP values, as estimated by the Clin model, for all risk biomarkers across all individuals. The clustering allowed us to uncover subgroups of individuals predicted to have high risk of OA. […] We present a large-scale study of OA, in a UK cohort of ~20,000 patients with OA and ~20,000 controls, encompassing a broad set of longitudinal OA risk biomarkers. We utilised interpretable machine learning to address gaps in the field of prediction and understanding of OA pathogenesis. Our models advance our understanding of risk of OA diagnosis, generating hypotheses for early screening, prevention and treatment of OA. The presented models predict an individuals 5-year risk of an OA diagnosis from EHR data which encompasses the range of OA heterogeneity and pathophysiology in real-life clinical settings. The complexity of OA risk was captured using multi-modal clinical data, biochemical and molecular signatures of OA. We identified distinct subgroups of OA risk profiles and derived simple clinical association rules for these subgroups. We also mapped differentially expressed molecular biomarkers between OA cases in risk subgroups. Finally, we demonstrated how individual patient journeys can be dissected to identify risk biomarkers for OA, contributing to the development of personalised preventative strategies.
#38 The KNee OsteoArthritis Prediction (KNOAP2020) challenge: An image analysis challenge to predict incident symptomatic radiographic knee osteoarthritis from MRI and X-ray images
https://pmc.ncbi.nlm.nih.gov/articles/PMC10323696/
The KNee OsteoArthritis Prediction (KNOAP2020) challenge was organized to objectively compare methods for the prediction of incident symptomatic radiographic knee osteoarthritis within 78 months on a test set with blinded ground truth. […] The KNOAP2020 challenge established a benchmark for predicting incident symptomatic radiographic knee osteoarthritis. Accurate prediction of incident symptomatic radiographic knee osteoarthritis is a complex and still unsolved problem requiring additional investigation. […] The aim of this challenge was to objectively compare different methods for the prediction of incident symptomatic radiographic knee OA (according to the combined American College of Rheumatology (ACR) criteria) within 78 months on a test set with blinded ground truth. […] The model with the highest ROC AUC (0.64) used a CNN-based model to extract information from X-ray images and combined that information with clinical variables (i.e., age, BMI, and KL grade). The model with the highest BACC (0.59) ensembled three different models that used automatically extracted X-ray and MRI features along with clinical variables. […] The performance of the submitted models on the independent test set with blinded ground truth was limited indicating that accurate prediction of incident symptomatic radiographic knee OA is a complex and still unsolved problem that requires additional investigation.
#39 PREDICTION MODELS TO ESTIMATE FUTURE INDIVIDUAL RISK OF OSTEOARTHRITIS IN THE GENERAL POPULATION: A SYSTEMATIC REVIEW
https://keele-repository.worktribe.com/output/507164/prediction-models-to-estimate-future-individual-risk-of-osteoarthritis-in-the-general-population-a-systematic-review
Conclusions: Of the 21 studies found and included in our review, 15 were published within the past 4 years, suggesting increasing interest among researchers in predicting individual-level risk for OA. […] However, models published to date remain heavily focussed on knee OA and have relied on a relatively small number of underlying cohort datasets. […] Furthermore, our systematic review highlights common shortfalls in applicability, suggesting that many models are not designed (nor yet intended) for mass application.
#40 The KNee OsteoArthritis Prediction (KNOAP2020) challenge: An image analysis challenge to predict incident symptomatic radiographic knee osteoarthritis from MRI and X-ray images
https://pmc.ncbi.nlm.nih.gov/articles/PMC10323696/
The KNee OsteoArthritis Prediction (KNOAP2020) challenge was organized to objectively compare methods for the prediction of incident symptomatic radiographic knee osteoarthritis within 78 months on a test set with blinded ground truth. […] The KNOAP2020 challenge established a benchmark for predicting incident symptomatic radiographic knee osteoarthritis. Accurate prediction of incident symptomatic radiographic knee osteoarthritis is a complex and still unsolved problem requiring additional investigation. […] The aim of this challenge was to objectively compare different methods for the prediction of incident symptomatic radiographic knee OA (according to the combined American College of Rheumatology (ACR) criteria) within 78 months on a test set with blinded ground truth. […] The model with the highest ROC AUC (0.64) used a CNN-based model to extract information from X-ray images and combined that information with clinical variables (i.e., age, BMI, and KL grade). The model with the highest BACC (0.59) ensembled three different models that used automatically extracted X-ray and MRI features along with clinical variables. […] The performance of the submitted models on the independent test set with blinded ground truth was limited indicating that accurate prediction of incident symptomatic radiographic knee OA is a complex and still unsolved problem that requires additional investigation.
#41 PREDICTION MODELS TO ESTIMATE FUTURE INDIVIDUAL RISK OF OSTEOARTHRITIS IN THE GENERAL POPULATION: A SYSTEMATIC REVIEW
https://keele-repository.worktribe.com/output/507164/prediction-models-to-estimate-future-individual-risk-of-osteoarthritis-in-the-general-population-a-systematic-review
Conclusions: Of the 21 studies found and included in our review, 15 were published within the past 4 years, suggesting increasing interest among researchers in predicting individual-level risk for OA. […] However, models published to date remain heavily focussed on knee OA and have relied on a relatively small number of underlying cohort datasets. […] Furthermore, our systematic review highlights common shortfalls in applicability, suggesting that many models are not designed (nor yet intended) for mass application.
#42 PREDICTION MODELS TO ESTIMATE FUTURE INDIVIDUAL RISK OF OSTEOARTHRITIS IN THE GENERAL POPULATION: A SYSTEMATIC REVIEW
https://keele-repository.worktribe.com/output/507164/prediction-models-to-estimate-future-individual-risk-of-osteoarthritis-in-the-general-population-a-systematic-review
Conclusions: Of the 21 studies found and included in our review, 15 were published within the past 4 years, suggesting increasing interest among researchers in predicting individual-level risk for OA. […] However, models published to date remain heavily focussed on knee OA and have relied on a relatively small number of underlying cohort datasets. […] Furthermore, our systematic review highlights common shortfalls in applicability, suggesting that many models are not designed (nor yet intended) for mass application.
#43
https://urncst.com/index.php/urncst/article/view/564
Knee Osteoarthritis (KOA) is the second most reported condition for persons 50 years and up; approximated by the continuous degradation of the knee, and eventually extending to the debilitation of biomechanical gait parameters. […] Of the 22 articles, 10% focused on the prediction of patient outcomes and disease prognosis, while 90% focused on the initial diagnosis or severity prediction of KOA. […] Overall, literature concerning the binary classification of symptomatic KOA provided high accuracy, yet further validation to minimize overfitting is required. Furthermore, areas for prognosis prediction and multiclass classification of KOA severity remain as areas for further development.
#44
https://urncst.com/index.php/urncst/article/view/564
Knee Osteoarthritis (KOA) is the second most reported condition for persons 50 years and up; approximated by the continuous degradation of the knee, and eventually extending to the debilitation of biomechanical gait parameters. […] Of the 22 articles, 10% focused on the prediction of patient outcomes and disease prognosis, while 90% focused on the initial diagnosis or severity prediction of KOA. […] Overall, literature concerning the binary classification of symptomatic KOA provided high accuracy, yet further validation to minimize overfitting is required. Furthermore, areas for prognosis prediction and multiclass classification of KOA severity remain as areas for further development.
#45 Prognostic model to predict the incidence of radiographic knee osteoarthritis | Annals of the Rheumatic Diseases
https://ard.bmj.com/content/83/5/661
Osteoarthritis Prognostic model to predict the incidence of radiographic knee osteoarthritis […] Early diagnosis of knee osteoarthritis (KOA) in asymptomatic stages is essential for the timely management of patients using preventative strategies. We develop and validate a prognostic model useful for predicting the incidence of radiographic KOA (rKOA) in non-radiographic osteoarthritic subjects and stratify individuals at high risk of developing the disease. […] A novel prognostic model based on common clinical variables and protein biomarkers was developed and externally validated to predict rKOA incidence over a 96-month period in individuals without any radiographic signs of disease. The resulting nomogram is a useful tool for stratifying high-risk populations and could potentially lead to personalised medicine strategies for treating OA.
#46 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
The Clin model confirmed that the biological and environmental risk factors underlying OA are heterogenous across individuals. We attempted to capture this heterogeneity and categorise patients into subgroups with differing risk biomarker profiles. Hence, we clustered the SHAP values, as estimated by the Clin model, for all risk biomarkers across all individuals. The clustering allowed us to uncover subgroups of individuals predicted to have high risk of OA. […] We present a large-scale study of OA, in a UK cohort of ~20,000 patients with OA and ~20,000 controls, encompassing a broad set of longitudinal OA risk biomarkers. We utilised interpretable machine learning to address gaps in the field of prediction and understanding of OA pathogenesis. Our models advance our understanding of risk of OA diagnosis, generating hypotheses for early screening, prevention and treatment of OA. The presented models predict an individuals 5-year risk of an OA diagnosis from EHR data which encompasses the range of OA heterogeneity and pathophysiology in real-life clinical settings. The complexity of OA risk was captured using multi-modal clinical data, biochemical and molecular signatures of OA. We identified distinct subgroups of OA risk profiles and derived simple clinical association rules for these subgroups. We also mapped differentially expressed molecular biomarkers between OA cases in risk subgroups. Finally, we demonstrated how individual patient journeys can be dissected to identify risk biomarkers for OA, contributing to the development of personalised preventative strategies.
#47 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
The Clin model confirmed that the biological and environmental risk factors underlying OA are heterogenous across individuals. We attempted to capture this heterogeneity and categorise patients into subgroups with differing risk biomarker profiles. Hence, we clustered the SHAP values, as estimated by the Clin model, for all risk biomarkers across all individuals. The clustering allowed us to uncover subgroups of individuals predicted to have high risk of OA. […] We present a large-scale study of OA, in a UK cohort of ~20,000 patients with OA and ~20,000 controls, encompassing a broad set of longitudinal OA risk biomarkers. We utilised interpretable machine learning to address gaps in the field of prediction and understanding of OA pathogenesis. Our models advance our understanding of risk of OA diagnosis, generating hypotheses for early screening, prevention and treatment of OA. The presented models predict an individuals 5-year risk of an OA diagnosis from EHR data which encompasses the range of OA heterogeneity and pathophysiology in real-life clinical settings. The complexity of OA risk was captured using multi-modal clinical data, biochemical and molecular signatures of OA. We identified distinct subgroups of OA risk profiles and derived simple clinical association rules for these subgroups. We also mapped differentially expressed molecular biomarkers between OA cases in risk subgroups. Finally, we demonstrated how individual patient journeys can be dissected to identify risk biomarkers for OA, contributing to the development of personalised preventative strategies.
#48 Prognostic model to predict the incidence of radiographic knee osteoarthritis | Annals of the Rheumatic Diseases
https://ard.bmj.com/content/83/5/661
A prognostic model based on clinical data and three protein biomarkers was developed and externally validated to predict the incidence of radiographic KOA in individuals without any radiographic signs of the disease, with an area under the curve of 0.83. […] This nomogram for KOA incidence is a useful tool for stratifying high-risk populations in order to prevent disease onset or delay its progression, and thereby decrease the associated burden. The prognostic model may also be valuable in improving patient recruitment into clinical trials.
#49 Prospective validity of a clinical prediction rule for response to non-surgical multidisciplinary management of knee osteoarthritis in tertiary care: a multisite prospective longitudinal study | BMJ Open
https://bmjopen.bmj.com/content/14/3/e078531
A clinical nomogram consisting of four simple measures to identify patients unlikely to respond to multidisciplinary non-surgical care for knee osteoarthritis in tertiary care has demonstrated capacity to predict poor response to non-surgical care, particularly when using higher cut-off scores. Further work is required to include a longer-term follow-up of participants in the current study, including both clinical and surgical outcomes.
#50 Prognostics for pain in osteoarthritis: Do clinical measures predict pain after total joint replacement? | PLOS One
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0222370
A significant proportion of osteoarthritis (OA) patients continue to experience moderate to severe pain after total joint replacement (TJR). […] Pre-surgical pain levels were not related to post-surgical residual pain. […] Our results show that although tested clinical and biopsychosocial variables reorganize after TJR in OA, their presurgical values are not predictive of post-surgery pain. […] Derivation of prognostic markers for pain persistence after TJR will require more comprehensive understanding of underlying mechanisms. […] Pain intensity prior to surgery, disproportion between pain intensity and articular damage, neuropathic-like symptoms, psychosocial factors such as pain catastrophizing and poor coping strategies are commonly referenced as important predictive factors. […] Our main aims are: 1) Test if baseline pain ratings relate to post-surgery pain levels; 2) examine how distinct pain measurement instruments relate to different clinical and biopsychological aspects of OA pain; 3) develop and evaluate models predictive of pain and pain relief after TJR; 4) use network analysis to assess the reorganization of pain related clinical and biopsychosocial properties of the personality of KOA and HOA patients after TJR.
#51 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
The Clin model confirmed that the biological and environmental risk factors underlying OA are heterogenous across individuals. We attempted to capture this heterogeneity and categorise patients into subgroups with differing risk biomarker profiles. Hence, we clustered the SHAP values, as estimated by the Clin model, for all risk biomarkers across all individuals. The clustering allowed us to uncover subgroups of individuals predicted to have high risk of OA. […] We present a large-scale study of OA, in a UK cohort of ~20,000 patients with OA and ~20,000 controls, encompassing a broad set of longitudinal OA risk biomarkers. We utilised interpretable machine learning to address gaps in the field of prediction and understanding of OA pathogenesis. Our models advance our understanding of risk of OA diagnosis, generating hypotheses for early screening, prevention and treatment of OA. The presented models predict an individuals 5-year risk of an OA diagnosis from EHR data which encompasses the range of OA heterogeneity and pathophysiology in real-life clinical settings. The complexity of OA risk was captured using multi-modal clinical data, biochemical and molecular signatures of OA. We identified distinct subgroups of OA risk profiles and derived simple clinical association rules for these subgroups. We also mapped differentially expressed molecular biomarkers between OA cases in risk subgroups. Finally, we demonstrated how individual patient journeys can be dissected to identify risk biomarkers for OA, contributing to the development of personalised preventative strategies.
#52 Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning | Nature Communications
https://www.nature.com/articles/s41467-024-46663-4
The Clin model confirmed that the biological and environmental risk factors underlying OA are heterogenous across individuals. We attempted to capture this heterogeneity and categorise patients into subgroups with differing risk biomarker profiles. Hence, we clustered the SHAP values, as estimated by the Clin model, for all risk biomarkers across all individuals. The clustering allowed us to uncover subgroups of individuals predicted to have high risk of OA. […] We present a large-scale study of OA, in a UK cohort of ~20,000 patients with OA and ~20,000 controls, encompassing a broad set of longitudinal OA risk biomarkers. We utilised interpretable machine learning to address gaps in the field of prediction and understanding of OA pathogenesis. Our models advance our understanding of risk of OA diagnosis, generating hypotheses for early screening, prevention and treatment of OA. The presented models predict an individuals 5-year risk of an OA diagnosis from EHR data which encompasses the range of OA heterogeneity and pathophysiology in real-life clinical settings. The complexity of OA risk was captured using multi-modal clinical data, biochemical and molecular signatures of OA. We identified distinct subgroups of OA risk profiles and derived simple clinical association rules for these subgroups. We also mapped differentially expressed molecular biomarkers between OA cases in risk subgroups. Finally, we demonstrated how individual patient journeys can be dissected to identify risk biomarkers for OA, contributing to the development of personalised preventative strategies.
#53 Prognostic model to predict the incidence of radiographic knee osteoarthritis – PubMed
https://pubmed.ncbi.nlm.nih.gov/38182405/
Objective: Early diagnosis of knee osteoarthritis (KOA) in asymptomatic stages is essential for the timely management of patients using preventative strategies. We develop and validate a prognostic model useful for predicting the incidence of radiographic KOA (rKOA) in non-radiographic osteoarthritic subjects and stratify individuals at high risk of developing the disease. […] Conclusion: A novel prognostic model based on common clinical variables and protein biomarkers was developed and externally validated to predict rKOA incidence over a 96-month period in individuals without any radiographic signs of disease. The resulting nomogram is a useful tool for stratifying high-risk populations and could potentially lead to personalised medicine strategies for treating OA.