A Dynamic Time Warping Extension to Consensus Weight-Based Cachexia Criteria Improves Prediction of Cancer Patient Outcomes

Abstract

Background: Cachexia is a complex syndrome that impacts up to half of patients with cancer. Criteria systems have been developed for the purpose of diagnosing and grading cachexia severity in clinical settings. One of the most widely known is those developed by Fearon et al. in 2011, which utilizes body mass loss and body mass index (BMI) to determine the presence and extent of cachexia. One limitation of this system and other clinical cachexia scales is the lack of systematic methods for assessing cachexia severity longitudinally. We sought to develop an extension to the 2011 consensus criteria that categorizes cancer patients with respect to their temporal cachexia progression and assess its predictive capacity relative to the current time-agnostic system.

Methods: Two cancer cohorts were identified in electronic health record data: lung cancer and glioblastoma. We extracted weight and BMI measures from the time of cancer diagnosis until death or loss to follow-up and computed cachexia severity according to the consensus criteria. Subgroups of cachexia progression were uncovered using dynamic time warping (DTW) followed by unsupervised clustering. This system and baseline consensus criteria measurements were each assessed for their ability to stratify patient outcomes utilizing Kaplan-Meier curves and Cox proportional hazards and subsequently compared with model concordance and inverse probability of censoring weighting (IPCW).

Results: Significant differences were observed in overall survival Kaplan-Meier curves of 1023 patients with lung cancer when stratified by baseline cachexia classification (p = 0.0002, N events = 592) but not in a cohort of 545 patients with glioblastoma (p = 0.16, N events = 353). DTW uncovered three patterns of cachexia progression in each subgroup with features described as 'smouldering', 'rapid with recovery' or 'persistent/recurrent'. Significant differences were observed in Kaplan-Meier curves when stratified by cachexia longitudinal patterns in lung cancer (p < 0.0001) and glioblastoma (p < 0.0001). Adjusted hazards ratios comparing the 'persistent/recurrent' cluster to referent subgroups in Cox models were 4.8 (4.1-5.8, p < 0.05) and 1.9 (1.4-2.4, p < 0.05) among patients with lung cancer and glioblastoma, respectively. Areas under the curve at multiple time points and Cox model concordances were greater when patients were stratified by progression pattern compared with baseline consensus criteria.

Conclusions: Our results suggest that accounting for cachexia's longitudinal progression in a systematic way can improve upon the prognostic capacity of a widely used consensus criteria set. These findings are important for the future development of systems that recognize concerning patterns of cachexia progression in clinical settings and aid clinicians in cachexia-related decision making.

Keywords: cancer cachexia; dynamic time warping; electronic health records.

FIGURE 1

Patient inclusion diagram. Left: Patients with lung cancer were initially included if a billed diagnosis of a lung malignancy was documented between 1 January 2012 and 31 December 2022 and if an order for administration of an immune checkpoint inhibitor occurred following the diagnosis. Right: Patients with GBM were identified by first querying the EDW for all patients with a billed diagnosis of a brain malignancy. Pathology documents were extracted from these patients that contained the word ‘glioblastoma’. A subset of these documents were reviewed for the presence of wild‐type IDH glioblastoma and a supervised machine learning algorithm was trained to identify patients matching this phenotype. Center: In both cohorts, patients were retained for analysis if the medical record contained a baseline weight measurement, defined as any measure made 6 months or less prior to diagnosis. To conduct the longitudinal analysis, patients were also required to have two or more weight measures following cancer diagnosis.

FIGURE 2

Baseline measurements of cachexia stratify outcomes in lung cancer but not in GBM. Kaplan–Meier curves were constructed, stratifying patients by their baseline measure. p values displayed in each panel were computed using the log‐rank test. DFS, disability‐free survival; HFS, hospitalization‐free survival; OS, overall survival.

FIGURE 3

Dynamic time warping and clustering on longitudinal consensus criteria measurements identifies patterns of cachexia progression. The percentage of patients still alive with criteria‐defined cachexia within each cluster was computed at 1‐week intervals from the time of cancer diagnosis for the following 3 years to illustrate the subgroup cachexia progression patterns. GBM, glioblastoma; Lung, lung cancer.

FIGURE 4

Cachexia progression patterns stratify patient outcomes in lung cancer and GBM. Kaplan–Meier curves of three patient‐centred outcomes among patients with lung cancer and GBM, stratified by their temporal cachexia progression pattern. p values were computed using the log‐rank test. DFS, disability‐free survival; HFS, hospitalization‐free survival; OS, overall survival.

FIGURE 5

Cachexia progression patterns stratify patient outcomes more effectively than baseline measures at multiple time points. Areas under the curve (AUC) for the baseline cachexia measure and temporal cachexia patterns were computed at 1‐month intervals starting at the time of cancer diagnosis until the median overall survival time was reached in each cohort. Values were compared between the baseline 2011 consensus criteria measure and temporal clusters to identify differences in cachexia measure performance in stratifying patient outcomes. The AUC describes the frequency with which the higher risk group—defined as cachexia for the baseline measurement and ‘persistent or recurrent’ for the temporal clusters—predict a shorter time to event than lower risk groups. Higher AUC values indicate variables that more effectively stratify patient outcomes. DFS, disability‐free survival; HFS, hospitalization‐free survival, OS; overall survival. Shaded region: confidence interval at the 95% confidence level. ‡Significant difference between receiver operating curves of baseline and temporal cluster Kaplan–Meier predictions at indicated time point and 95% confidence level.