A comparison of different population-level summary measures for randomised trials with time-to-event outcomes, with a focus on non-inferiority trials.

BACKGROUND: The population-level summary measure is a key component of the estimand for clinical trials with time-to-event outcomes. This is particularly the case for non-inferiority trials, because different summary measures imply different null hypotheses. Most trials are designed using the hazard ratio as summary measure, but recent studies suggested that the difference in restricted mean survival time might be more powerful, at least in certain situations. In a recent letter, we conjectured that differences between summary measures can be explained using the concept of the non-inferiority frontier and that for a fair simulation comparison of summary measures, the same analysis methods, making the same assumptions, should be used to estimate different summary measures. The aim of this article is to make such a comparison between three commonly used summary measures: hazard ratio, difference in restricted mean survival time and difference in survival at a fixed time point. In addition, we aim to investigate the impact of using an analysis method that assumes proportional hazards on the operating characteristics of a trial designed with any of the three summary measures. METHODS: We conduct a simulation study in the proportional hazards setting. We estimate difference in restricted mean survival time and difference in survival non-parametrically, without assuming proportional hazards. We also estimate all three measures parametrically, using flexible survival regression, under the proportional hazards assumption. RESULTS: Comparing the hazard ratio assuming proportional hazards with the other summary measures not assuming proportional hazards, relative performance varies substantially depending on the specific scenario. Fixing the summary measure, assuming proportional hazards always leads to substantial power gains compared to using non-parametric methods. Fixing the modelling approach to flexible parametric regression assuming proportional hazards, difference in restricted mean survival time is most often the most powerful summary measure among those considered. CONCLUSION: When the hazards are likely to be approximately proportional, reflecting this in the analysis can lead to large gains in power for difference in restricted mean survival time and difference in survival. The choice of summary measure for a non-inferiority trial with time-to-event outcomes should be made on clinical grounds; when any of the three summary measures discussed here is equally justifiable, difference in restricted mean survival time is most often associated with the most powerful test, on the condition that it is estimated under proportional hazards.

Increasing frequency of gene copy number aberrations is associated with immunosuppression and predicts poor prognosis in gastric adenocarcinoma.

BACKGROUND: Patients with Epstein-Barr virus-positive gastric cancers or those with microsatellite instability appear to have a favourable prognosis. However, the prognostic value of the chromosomal status (chromosome-stable (CS) versus chromosomal instable (CIN)) remains unclear in gastric cancer. METHODS: Gene copy number aberrations (CNAs) were determined in 16 CIN-associated genes in a retrospective study including test and validation cohorts of patients with gastric cancer. Patients were stratified into CS (no CNA), CINlow (1-2 CNAs) or CINhigh (3 or more CNAs). The relationship between chromosomal status, clinicopathological variables, and overall survival (OS) was analysed. The relationship between chromosomal status, p53 expression, and tumour infiltrating immune cells was also assessed and validated externally. RESULTS: The test and validation cohorts included 206 and 748 patients, respectively. CINlow and CINhigh were seen in 35.0 and 15.0 per cent of patients, respectively, in the test cohort, and 48.5 and 20.7 per cent in the validation cohort. Patients with CINhigh gastric cancer had the poorest OS in the test and validation cohorts. In multivariable analysis, CINlow, CINhigh and pTNM stage III-IV (P < 0.001) were independently associated with poor OS. CIN was associated with high p53 expression and low immune cell infiltration. CONCLUSION: CIN may be a potential new prognostic biomarker independent of pTNM stage in gastric cancer. Patients with gastric cancer demonstrating CIN appear to be immunosuppressed, which might represent one of the underlying mechanisms explaining the poor survival and may help guide future therapeutic decisions.

Automated detection and delineation of lymph nodes in haematoxylin & eosin stained digitised slides.

Treatment of patients with oesophageal and gastric cancer (OeGC) is guided by disease stage, patient performance status and preferences. Lymph node (LN) status is one of the strongest prognostic factors for OeGC patients. However, survival varies between patients with the same disease stage and LN status. We recently showed that LN size from patients with OeGC might also have prognostic value, thus making delineations of LNs essential for size estimation and the extraction of other imaging biomarkers. We hypothesized that a machine learning workflow is able to: (1) find digital H&E stained slides containing LNs, (2) create a scoring system providing degrees of certainty for the results, and (3) delineate LNs in those images. To train and validate the pipeline, we used 1695 H&E slides from the OE02 trial. The dataset was divided into training (80%) and validation (20%). The model was tested on an external dataset of 826 H&E slides from the OE05 trial. U-Net architecture was used to generate prediction maps from which predefined features were extracted. These features were subsequently used to train an XGBoost model to determine if a region truly contained a LN. With our innovative method, the balanced accuracies of the LN detection were 0.93 on the validation dataset (0.83 on the test dataset) compared to 0.81 (0.81) on the validation (test) datasets when using the standard method of thresholding U-Net predictions to arrive at a binary mask. Our method allowed for the creation of an "uncertain" category, and partly limited false-positive predictions on the external dataset. The mean Dice score was 0.73 (0.60) per-image and 0.66 (0.48) per-LN for the validation (test) datasets. Our pipeline detects images with LNs more accurately than conventional methods, and high-throughput delineation of LNs can facilitate future LN content analyses of large datasets.

Neoadjuvant chemotherapy improves survival in patients with oesophageal mucinous adenocarcinoma: Post-hoc analysis of the UK MRC OE02 and OE05 trials.

BACKGROUND: Adenocarcinoma with more than 50% extracellular mucin is a relatively rare histological subtype of gastrointestinal adenocarcinomas. The clinical impact of extracellular mucin in oesophageal adenocarcinoma (OeAC) has not been investigated in detail. We hypothesised that patients with mucinous OeAC (OeACmucin) do not benefit from neoadjuvant chemotherapy. METHODS: OeAC patients either treated by surgery alone in the OE02 trial (S-patients) or by neoadjuvant chemotherapy followed by surgery (CS-patients) in OE02 or OE05 trials were included. Cancers from 1055 resection specimens (OE02 [test cohort]: 187 CS, 185 S; OE05 [validation cohort]: 683 CS) were classified as either mucinous (more than 50% of the tumour area consists of extracellular mucin, OeACmucin) or non-mucinous adenocarcinoma (OeACnon-mucin). The relationship between histological phenotype, clinicopathological characteristics, survival and treatment was analysed. RESULTS: Overall, 7.3% and 9.6% OeAC were classified as OeACmucin in OE02 and OE05, respectively. In OE02, the frequency of OeACmucin was similar in S and CS-patients. Patients with OeACmucin treated with surgery alone had a poorer overall survival compared with OeACnon-mucin patients (hazard ratio: 2.222, 95% confidence interval: 1.08-4.56, P = 0.025). Patients with OeACmucin treated with neoadjuvant chemotherapy and surgery had similar survival as OeACnon-mucin patients in test and validation cohort. CONCLUSIONS: This is the first study to suggest in a post-hoc analysis of material from two independent phase III clinical trials that the poor survival of patients with mucinous OeAC can be improved by neoadjuvant chemotherapy. Future studies are warranted to identify potential underlying biological, biochemical or pharmacokinetic interactions between extracellular mucin and chemotherapy.

Development and validation of deep learning classifiers to detect Epstein-Barr virus and microsatellite instability status in gastric cancer: a retrospective multicentre cohort study.

BACKGROUND: Response to immunotherapy in gastric cancer is associated with microsatellite instability (or mismatch repair deficiency) and Epstein-Barr virus (EBV) positivity. We therefore aimed to develop and validate deep learning-based classifiers to detect microsatellite instability and EBV status from routine histology slides. METHODS: In this retrospective, multicentre study, we collected tissue samples from ten cohorts of patients with gastric cancer from seven countries (South Korea, Switzerland, Japan, Italy, Germany, the UK and the USA). We trained a deep learning-based classifier to detect microsatellite instability and EBV positivity from digitised, haematoxylin and eosin stained resection slides without annotating tumour containing regions. The performance of the classifier was assessed by within-cohort cross-validation in all ten cohorts and by external validation, for which we split the cohorts into a five-cohort training dataset and a five-cohort test dataset. We measured the area under the receiver operating curve (AUROC) for detection of microsatellite instability and EBV status. Microsatellite instability and EBV status were determined to be detectable if the lower bound of the 95% CI for the AUROC was above 0·5. FINDINGS: Across the ten cohorts, our analysis included 2823 patients with known microsatellite instability status and 2685 patients with known EBV status. In the within-cohort cross-validation, the deep learning-based classifier could detect microsatellite instability status in nine of ten cohorts, with AUROCs ranging from 0·597 (95% CI 0·522-0·737) to 0·836 (0·795-0·880) and EBV status in five of eight cohorts, with AUROCs ranging from 0·819 (0·752-0·841) to 0·897 (0·513-0·966). Training a classifier on the pooled training dataset and testing it on the five remaining cohorts resulted in high classification performance with AUROCs ranging from 0·723 (95% CI 0·676-0·794) to 0·863 (0·747-0·969) for detection of microsatellite instability and from 0·672 (0·403-0·989) to 0·859 (0·823-0·919) for detection of EBV status. INTERPRETATION: Classifiers became increasingly robust when trained on pooled cohorts. After prospective validation, this deep learning-based tissue classification system could be used as an inexpensive predictive biomarker for immunotherapy in gastric cancer. FUNDING: German Cancer Aid and German Federal Ministry of Health.

DNA epigenetic signature predictive of benefit from neoadjuvant chemotherapy in oesophageal adenocarcinoma: results from the MRC OE02 trial.

BACKGROUND: DNA methylation signatures describing distinct histological subtypes of oesophageal cancer have been reported. We studied DNA methylation in samples from the MRC OE02 phase III trial, which randomised patients with resectable oesophageal cancer to surgery alone (S) or neoadjuvant chemotherapy followed by surgery (CS). AIM: The aim of the study was to identify epigenetic signatures predictive of chemotherapy benefit in patients with oesophageal adenocarcinoma (OAC) from the OE02 trial and validate the findings in an independent cohort. METHODS: DNA methylation was analysed using the Illumina GoldenGate platform on surgically resected OAC specimens from patients in the OE02 trial. Cox proportional hazard analysis was performed to select probes predictive of survival in the CS arm. Non-negative matrix factorisation was used to perform clustering and delineate DNA methylation signatures. The findings were validated in an independent cohort of patients with gastroesophageal adenocarcinoma treated with neoadjuvant chemotherapy. RESULTS: A total of 229 patients with OAC were analysed from the OE02 trial (118 in the CS arm and 111 in the S arm). There was no difference in DNA methylation status between the CS and S arms. A metagene signature was created by dichotomising samples into two clusters. In cluster 1, patients in the CS arm had significant overall survival (OS) benefit (median OS CS: 931 days vs. S: 536 days [HR: 1.54, P = 0.031]). In cluster 2, patients in the CS arm had similar (or worse) OS compared with patients in the S arm (CS: 348 days vs. S: 472 days [HR: 0.70, P = 0.1], and test of interaction was significant (p = 0.005). In the validation cohort (n = 13), there was no difference in DNA methylation status in paired pre- and post-treatment samples. When the epigenetic signature was applied, cluster 1 samples had better OS (median OS, cluster 1: 1174 days vs. cluster 2: 392 days, HR: 3.47, p = 0.059) CONCLUSIONS: This is the first and largest study of DNA methylation in patients with OAC uniformly treated in a randomised phase III trial. We identified an epigenetic signature that may serve as a predictive biomarker for chemotherapy benefit in OAC.