Does a Concise Patient-reported Outcome Measure Provide a Valid Measure of Physical Function for Cancer Patients...

Current functional assessment tools for orthopaedic oncology are long surveys that contribute to patients' survey fatigue and yet lack the ability to discern meaningful differences in a patient population that is often mobile but unable to perform strenuous activities. We sought to determine whether a shorter, novel tool based on existing, validated surveys could better capture differences in a sample of orthopaedic oncology patients.

(1) Can a concise fixed-item functional tool derived from the 50 items in the Toronto Extremity Salvage Score for the lower extremity (TESS LE) and the Lower Extremity Functional Scale (LEFS) demonstrate similar responsiveness in terms of sensitivity and specificity? (2) What is the precision and accuracy of the concise tool compared with the TESS LE and LEFS?

Functional outcome data were collected and maintained in a longitudinally maintained database at a single institution. Patients were included in the study if (1) they had undergone a tumor excision or a nononcologic orthopaedic procedure (for example, arthroplasty for osteoarthritis) for a bone or soft tissue tumor affecting lower extremity function, and (2) they had completed the LEFS, TESS LE, and Patient-Reported Outcomes Measurement Information System (PROMIS) global health tool on at least two clinic visits. Between September 2014 and April 2022, we treated 14,234 patients for primary bone or soft tissue sarcoma, metastatic disease to bone, or orthopaedic sequelae of chronic cancer care. Approximately 6% (854 of 14,234) were excluded due to the need of a language translator. Approximately 2% (278 of 13,380) refused or were unable to participate. Seventy-two percent (9433 of 13,102) of the patients had an operation on a lower extremity. Of these, 4% (339 of 9433) of the patients completed the TESS LE, LEFS, and Item 3 of the PROMIS global health tool on ≥ 2 clinic visits. Of the patients in the current study, 49% (167 of 339) were women, and 27% (93 of 339) had metastatic carcinoma. Twelve percent (41 of 339) of the patients died before the end of the study period. Spearman rank-order correlation, principal component analysis (PCA), and item response theory (IRT) modeling identified 14 highly discriminating items from the TESS LE and LEFS. Multiple linear stepwise regression (MLSR) was performed with the dependent variable being the summary score of the 14 items derived from the TESS LE and LEFS and standardized to a percentage of 100. The beta coefficient from the MLSR was used to derive a weight for each of the 14 items. Evaluation of the model with 10 to 17 variables was performed to ensure that the model with the 14 items met the most criteria for fit with the PCA, the receiver operating characteristic (ROC) curve, and the IRT modeling criteria. The responsiveness (sensitivity and specificity) of the change scores in the shortened 14-item survey, the 30-item TESS LE, and the 20-item LEFS as compared with the dichotomized changes in Item 3 of the PROMIS global health tool was evaluated using ROCs. The concordance (accuracy and precision) of the 14 items derived from the LEFS and TESS LE was evaluated.

The responsiveness (sensitivity and specificity) of the shortened 14-item survey, the TESS LE, and the LEFS to the criterion target of the PROMIS global health tool (Item 3) was similar, with areas under the curve (AUCs) ranging from 0.62 to 0.65 for the ROC curves. The responsiveness of the 14-item survey to the TESS LE showed sensitivity of 96% and specificity of 90%, with an AUC of 0.98 (p < 0.001). The responsiveness of the 14 items to the LEFS showed sensitivity of 95% and specificity of 86%, with an AUC of 0.96. The validity of the 14 items to the TESS LE was measured by concordance, with a precision of 0.98 and an accuracy of 0.97. Concordance of the 14 items to the LEFS showed a precision of 0.98 and accuracy of 0.83.

The concise 14 items derived from patient-reported responses in the TESS LE and LEFS outcome measures showed similar responsiveness (sensitivity and specificity) as the original TESS LE and LEFS for cancer patients after lower extremity orthopaedic surgery performed for oncologic and nononcologic indications. The concise 14 items have a similar ability to the TESS LE and LEFS to tell the clinician or patient how they are functioning compared with other patients. These 14 items are shorter than the combined 50 items of the TESS LE and LEFS while retaining the capacity to describe a broad range of lower extremity function for orthopaedic oncology patients. We have named the 14-item survey the Lower Extremity Oncology Functional Assessment Tool (LEO).Level of Evidence Level II, diagnostic study.