Negative prognostic factors and clinical improvement prediction modeling for extracorporeal shockwave therapy in calcific shoulder tendinitis using artificial intelligence techniques

Comprehensive Summary

This study examines whether the outcomes of extracorporeal shockwave therapy (ESWT) can be predicted by artificial intelligence (AI) in patients with calcific shoulder tendinitis, a leading cause of shoulder pain and commonly resulting from calcium buildup in the rotator and cuff tendons. The researchers retrospectively analyzed the data of 296 patients who underwent ESWT between 1998 and 2022 after other conservative treatments failed. To conduct the study, they trained an eXtreme Gradient Boosting (XGBoost) machine learning model with patient-specific factors: age, symptom duration, calcification size and pre-treatment pain and function scores. This was used to predict whether ESWT would improve pain (measured by VAS score) and function (measured by Constant-Murley score). Consequently, they identified three key negative prognostic factors of ESWT failure, including symptom duration longer than 10 months, severe pain before treatment (VAS > 5) and higher pre-treatment functional scores (CMS > 55). Additionally, the XGBoost model had high accuracy, even when only 4 variables were used. This suggests a strong potential for clinical prediction of ESWT outcomes.

Outcomes and Implications

The findings of this study are clinically significant as they help to identify whether patients are likely to benefit from ESWT. This may in the future allow physicians to recommend alternative treatments earlier, limiting medical costs, mental distress and patient frustrations. Medical conditions such as calcific tendinitis can significantly impact patients' quality of life and ability to work. An early understanding that they may not respond ESWT could improve the outcomes and reduce unnecessary surgeries or procedures. Although the model displayed high accuracy in its predictions, further research is needed before wider implementations in clinical settings.