The ROC curve for regularly measured longitudinal biomarkers

Abstract

The receiver operating characteristic (ROC) curve is a commonly used graphical summary of the discriminative capacity of a thresholded continuous scoring system for a binary outcome. Estimation and inference procedures for the ROC curve are well-studied in the cross-sectional setting. However, there is a paucity of research when both biomarker measurements and disease status are observed longitudinally. In a motivating example, we are interested in characterizing the value of longitudinally measured CD4 counts for predicting the presence or absence of a transient spike in HIV viral load, also time-dependent. The existing method neither appropriately characterizes the diagnostic value of observed CD4 counts nor efficiently uses status history in predicting the current spike status. We propose to jointly model the binary status as a Markov chain and the biomarkers levels, conditional on the binary status, as an autoregressive process, yielding a dynamic scoring procedure for predicting the occurrence of a spike. Based on the resulting prediction rule, we propose several natural extensions of the ROC curve to the longitudinal setting and describe procedures for statistical inference. Lastly, extensive simulations have been conducted to examine the small sample operational characteristics of the proposed methods.

Combining total and differential somatic cell count to screen for mastitis

“…Nevertheless, cut-off value of the investigated marker (our major objective) could not be provided in the output of the SAS macro. Moreover, as mentioned by Michael et al (2019), such approach does not account for the possible greater dependency between marker and binary outcome nearer in time (in our case between data measured in subsequent test-day records). To address these issues, Michael et al (2019) proposed prediction rules with modelling based on Markov chain and autoregressive process to extend the ROC procedure to longitudinal studies.…”

“…Moreover, as mentioned by Michael et al (2019), such approach does not account for the possible greater dependency between marker and binary outcome nearer in time (in our case between data measured in subsequent test-day records). To address these issues, Michael et al (2019) proposed prediction rules with modelling based on Markov chain and autoregressive process to extend the ROC procedure to longitudinal studies. As mentioned by the authors, this approach was developed assuming that the marker was measured at regular time intervals (and this is not the case of milk recording system) and appropriate model checking analysis to adapt it to our data was not feasible.…”

An Immune-Related Signature for Predicting the Prognosis of Lower-Grade Gliomas

“…With the R package “survminer”, Kaplan–Meier plots and the log-rank test were used to estimate the survival rate between the low- and high-risk groups ( 14 ). A time-dependent receiver operating characteristic (ROC) curve was calculated to assess the predictive value of the multivariate Cox model ( 15 , 16 ). To rule out the factors that cause accidental death in patients, such as death from postoperative complications, we excluded samples with a follow-up or OS time shorter than 90 days.…”

Integration of Face-to-Face Screening With Real-time Machine Learning to Predict Risk of Suicide Among Adults

“…Since patients with more than 1 visit were included multiple times, discrimination metrics might have been affected by the repeatedobservations problem. 39 By defining the outcomes of interest (SA or SI) based on ICD codes associated with follow-up visits at VUMC, we excluded cases occurring outside our health care system. ICD codes may have introduced diagnostic imprecision or excluded incompletely coded cases of suicide (eg, patients presenting in cardiac arrest or in overdose without obvious suicidal association).…”