In oncology clinical trials, tumor burden (TB) stands as a crucial longitudinal biomarker, reflecting the toll a tumor takes on a patient's prognosis. With certain treatments, the disease's natural progression shows the tumor burden initially receding before rising once more. Biologically, the point of change may be different between individuals and must have occurred between the baseline measurement and progression time of the patient, implying a random effects model obeying a bound constraint. However, in practice, patients may drop out of the study due to progression or death, presenting a non-ignorable missing data problem. In this paper, we introduce a novel joint model that combines time-to-event data and longitudinal data, where the latter is parameterized by a random change point augmented by random pre-slope and post-slope dynamics. Importantly, the model is equipped to incorporate covariates across the longitudinal and survival models, adding significant flexibility. Adopting a Bayesian approach, we propose an efficient Hamiltonian Monte Carlo (HMC) algorithm for parameter inference. We demonstrate the superiority of our approach compared to a longitudinal-only model via simulations and apply our method to a data set in oncology. The code for implementation is publicly available on https://github.com/quyixiang/chgptModel.

Therapy decision for patients with metastatic renal cell carcinoma (mRCC) is highly dependent on disease monitoring based on radiological reports. The purpose of the study was to compare non-standardized, common practice free text reporting (FTR) on disease response with reporting based on response evaluation criteria in solid tumors modified for immune-based therapeutics (iRECIST).

Fifty patients with advanced mRCC were included in the retrospective, single-center study. CT scans had been evaluated and FTR prepared in accordance with center's routine practice. For study purposes, reports were re-evaluated using a dedicated computer program that applied iRECIST. Patients were followed up over a period of 22.8 ± 7.9 months in intervals of 2.7 ± 1.8 months. Weighted kappa statistics was run to assess strength of agreement. Logistic regression was used to identify predictors for different rating.

Agreement between FTR and iRECIST-based reporting was moderate (kappa 0.38 [95% CI 0.2-0.6] to 0.70 [95% CI 0.5-0.9]). Tumor response or progression according to FTR were not confirmed with iRECIST in 19 (38%) or 11 (22%) patients, respectively, in at least one follow-up examination. With FTR, new lesions were frequently not recognized if they were already identified in the recent prior follow-up examination (odds ratio for too favorable rating of disease response compared to iRECIST: 5.4 [95% CI 2.9-10.1].

Moderate agreement between disease response according to FTR or iRECIST in patients with mRCC suggests the need of standardized quantitative radiological assessment in daily clinical practice.

Nonlinear joint models are a powerful tool to precisely analyze the association between a nonlinear biomarker and a time-to-event process, such as death. Here, we review the main methodological techniques required to build these models and to make inferences and predictions. We describe the main clinical applications and discuss the future developments of such models.

Recent studies have discovered several unique tumor response subgroups outside of response classification by Response Evaluation Criteria for Solid Tumors (RECIST), such as mixed response and oligometastasis. These subtypes have a distinctive property, lesion heterogeneity defined as diversity of tumor growth profiles in RECIST target lesions. Furthermore, many cancer clinical trials have been activated to evaluate various treatment options for heterogeneity-related subgroups (e.g., 29 trials so far listed in clinicaltrials.gov for cancer patients with oligometastasis). Some of the trials have shown survival benefit by tailored treatment strategies. This evidence presents the unmet need to incorporate lesion heterogeneity to improve RECIST response classification.

An approach for Lesion Heterogeneity Classification (LeHeC) was developed using a contemporary statistical approach to assess target lesion variation, characterize patient treatment response, and translate informative evidence to improving treatment strategy. A mixed effect linear model was used to determine lesion heterogeneity. Further analysis was conducted to classify various types of lesion variation and incorporate with RECIST to enhance response classification. A study cohort of 110 target lesions from 36 lung cancer patients was used for evaluation.

Due to small sample size issue, the result was exploratory in nature. By analyzing RECIST target lesion data, the LeHeC approach detected a high prevalence (n = 21; 58%) of lesion heterogeneity. Subgroup classification revealed several informative distinct subsets in a descending order of lesion heterogeneity: mix of progression and regression (n = 7), mix of progression and stability (n = 9), mix of regression and stability (n = 5), and non-heterogeneity (n = 15). Evaluation for association of lesion heterogeneity and RECIST best response classification showed lesion heterogeneity commonly occurred in each response group (stable disease: 16/27; 59%; partial response: 3/5; 60%; progression disease: 2/4; 50%). Survival analysis showed a differential trend of overall survival between heterogeneity and non-heterogeneity in RECIST response groups.

This is the first study to evaluate lesion heterogeneity, an underappreciated metric, for RECIST application in oncology clinical trials. Results indicated lesion heterogeneity is not an uncommon event. The LeHeC approach could enhance RECIST response classification by utilizing granular lesion level discovery of heterogeneity.

The tumor objective response rate (ORR) is an important parameter to demonstrate the efficacy of a treatment in oncology. The ORR is valuable for clinical decision making in routine practice and a significant end-point for reporting the results of clinical trials. World Health Organization and Response Evaluation Criteria in Solid Tumors (RECIST) are anatomic response criteria developed mainly for cytotoxic chemotherapy. These criteria are based on the visual assessment of tumor size in morphological images provided by computed tomography (CT) or magnetic resonance imaging. Anatomic response criteria may not be optimal for biologic agents, some disease sites, and some regional therapies. Consequently, modifications of RECIST, Choi criteria and Morphologic response criteria were developed based on the concept of the evaluation of viable tumors. Despite its limitations, RECIST v1.1 is validated in prospective studies, is widely accepted by regulatory agencies and has recently shown good performance for targeted cancer agents. Finally, some alternatives of RECIST were developed as immune-specific response criteria for checkpoint inhibitors. Immune RECIST criteria are based essentially on defining true progressive disease after a confirmatory imaging. Some graphical methods may be useful to show longitudinal change in the tumor burden over time. Tumor tissue is a tridimensional heterogenous mass, and tumor shrinkage is not always symmetrical; thus, metabolic response assessments using positron emission tomography (PET) or PET/CT may reflect the viability of cancer cells or functional changes evolving after anticancer treatments. The metabolic response can show the benefit of a treatment earlier than anatomic shrinkage, possibly preventing delays in drug approval. Computer-assisted automated volumetric assessments, quantitative multimodality imaging in radiology, new tracers in nuclear medicine and finally artificial intelligence have great potential in future evaluations.

Fluoro-L-thymidine (FLT) is a positron emission tomography/computed tomography (PET/CT) tracer which reflects proliferative activity in a cancer lesion. The main objective of this prospective explorative study was to evaluate whether FLT-PET can be used for the early evaluation of treatment response in colorectal cancer patients (CRC) with liver metastases. Patients with metastatic CRC having at least one measurable (>1 cm) liver metastasis receiving first-line chemotherapy were included. A FLT-PET/CT scan was performed at baseline and after the first treatment. The maximum and mean standardised uptake values (SUVmax, SUVmean) were measured. After three cycles of chemotherapy, treatment response was assessed by CT scan based on RECIST 1.1.

Thirty-nine consecutive patients were included of which 27 were evaluable. Dropout was mainly due to disease complications. Nineteen patients (70%) had a partial response, seven (26%) had stable disease and one (4%) had progressive disease. A total of 23 patients (85%) had a decrease in FLT uptake following the first treatment. The patient with progressive disease had the highest increase in FLT uptake in SUVmax. There was no correlation between the response according to RECIST and the early changes in FLT uptake measured as SUVmax (p = 0.24).

No correlation was found between early changes in FLT uptake after the first cycle of treatment and the response evaluated from subsequent CT scans. It seems unlikely that FLT-PET can be used on its own for the early response evaluation of metastatic CRC.

Joint models for longitudinal and time-to-event data are commonly used to simultaneously analyse correlated data in single study cases. Synthesis of evidence from multiple studies using meta-analysis is a natural next step but its feasibility depends heavily on the standard of reporting of joint models in the medical literature. During this review we aim to assess the current standard of reporting of joint models applied in the literature, and to determine whether current reporting standards would allow or hinder future aggregate data meta-analyses of model results.

We undertook a literature review of non-methodological studies that involved joint modelling of longitudinal and time-to-event medical data. Study characteristics were extracted and an assessment of whether separate meta-analyses for longitudinal, time-to-event and association parameters were possible was made.

The 65 studies identified used a wide range of joint modelling methods in a selection of software. Identified studies concerned a variety of disease areas. The majority of studies reported adequate information to conduct a meta-analysis (67.7% for longitudinal parameter aggregate data meta-analysis, 69.2% for time-to-event parameter aggregate data meta-analysis, 76.9% for association parameter aggregate data meta-analysis). In some cases model structure was difficult to ascertain from the published reports.

Whilst extraction of sufficient information to permit meta-analyses was possible in a majority of cases, the standard of reporting of joint models should be maintained and improved. Recommendations for future practice include clear statement of model structure, of values of estimated parameters, of software used and of statistical methods applied.