Validation of Predictive Analyses for Interim Decisions in Clinical Trials

Risk stratification in gastric cancer lung metastasis: Utilizing an overall survival nomogram and comparing it with previous staging

BACKGROUND

Gastric cancer (GC) is prevalent and aggressive, especially when patients have distant lung metastases, which often places patients into advanced stages. By identifying prognostic variables for lung metastasis in GC patients, it may be possible to construct a good prediction model for both overall survival (OS) and the cumulative incidence prediction (CIP) plot of the tumour.

AIM

To investigate the predictors of GC with lung metastasis (GCLM) to produce nomograms for OS and generate CIP by using cancer-specific survival (CSS) data.

METHODS

Data from January 2000 to December 2020 involving 1652 patients with GCLM were obtained from the Surveillance, epidemiology, and end results program database. The major observational endpoint was OS; hence, patients were separated into training and validation groups. Correlation analysis determined various connections. Univariate and multivariate Cox analyses validated the independent predictive factors. Nomogram distinction and calibration were performed with the time-dependent area under the curve (AUC) and calibration curves. To evaluate the accuracy and clinical usefulness of the nomograms, decision curve analysis (DCA) was performed. The clinical utility of the novel prognostic model was compared to that of the 7th edition of the American Joint Committee on Cancer (AJCC) staging system by utilizing Net Reclassification Improvement (NRI) and Integrated Discrimination Improvement (IDI). Finally, the OS prognostic model and Cox-AJCC risk stratification model modified for the AJCC system were compared.

RESULTS

For the purpose of creating the OS nomogram, a CIP plot based on CSS was generated. Cox multivariate regression analysis identified eleven significant prognostic factors (P < 0.05) related to liver metastasis, bone metastasis, primary site, surgery, regional surgery, treatment sequence, chemotherapy, radiotherapy, positive lymph node count, N staging, and time from diagnosis to treatment. It was clear from the DCA (net benefit > 0), time-dependent ROC curve (training/validation set AUC > 0.7), and calibration curve (reliability slope closer to 45 degrees) results that the OS nomogram demonstrated a high level of predictive efficiency. The OS prediction model (New Model AUC = 0.83) also performed much better than the old Cox-AJCC model (AUC difference between the new model and the old model greater than 0) in terms of risk stratification (P < 0.0001) and verification using the IDI and NRI.

CONCLUSION

The OS nomogram for GCLM successfully predicts 1- and 3-year OS. Moreover, this approach can help to appropriately classify patients into high-risk and low-risk groups, thereby guiding treatment.

Data sources and applied methods for paclitaxel safety signal discernment

Background

Following the identification of a late mortality signal, the Food and Drug Administration (FDA) convened an advisory panel that concluded that additional clinical study data are needed to comprehensively evaluate the late mortality signal observed with the use of drug-coated balloons (DCB) and drug-eluting stent (DES). The objective of this review is to (1) identify and summarize the existing clinical and cohort studies assessing paclitaxel-coated DCBs and DESs, (2) describe and determine the quality of the available data sources for the evaluation of these devices, and (3) present methodologies that can be leveraged for proper signal discernment within available data sources.

Methods

Studies and data sources were identified through comprehensive searches. original research studies, clinical trials, comparative studies, multicenter studies, and observational cohort studies written in the English language and published from January 2007 to November 2021, with a follow-up longer than 36 months, were included in the review. Data quality of available data sources identified was assessed in three groupings. Moreover, accepted data-driven methodologies that may help circumvent the limitations of the extracted studies and data sources were extracted and described.

Results

There were 39 studies and data sources identified. This included 19 randomized clinical trials, nine single-arm studies, eight registries, three administrative claims, and electronic health records. Methodologies focusing on the use of existing premarket clinical data, the incorporation of all contributed patient time, the use of aggregated data, approaches for individual-level data, machine learning and artificial intelligence approaches, Bayesian approaches, and the combination of various datasets were summarized.

Conclusion

Despite the multitude of available studies over the course of eleven years following the first clinical trial, the FDA-convened advisory panel found them insufficient for comprehensively assessing the late-mortality signal. High-quality data sources with the capabilities of employing advanced statistical methodologies are needed to detect potential safety signals in a timely manner and allow regulatory bodies to act quickly when a safety signal is detected.