A CT-based radiomics classification model for the prediction of histological type and tumour grade in retroperitoneal sarcoma (RADSARC-R): a retrospective multicohort analysis.

BACKGROUND: Retroperitoneal sarcomas are tumours with a poor prognosis. Upfront characterisation of the tumour is difficult, and under-grading is common. Radiomics has the potential to non-invasively characterise the so-called radiological phenotype of tumours. We aimed to develop and independently validate a CT-based radiomics classification model for the prediction of histological type and grade in retroperitoneal leiomyosarcoma and liposarcoma. METHODS: A retrospective discovery cohort was collated at our centre (Royal Marsden Hospital, London, UK) and an independent validation cohort comprising patients recruited in the phase 3 STRASS study of neoadjuvant radiotherapy in retroperitoneal sarcoma. Patients aged older than 18 years with confirmed primary leiomyosarcoma or liposarcoma proceeding to surgical resection with available contrast-enhanced CT scans were included. Using the discovery dataset, a CT-based radiomics workflow was developed, including manual delineation, sub-segmentation, feature extraction, and predictive model building. Separate probabilistic classifiers for the prediction of histological type and low versus intermediate or high grade tumour types were built and tested. Independent validation was then performed. The primary objective of the study was to develop radiomic classification models for the prediction of retroperitoneal leiomyosarcoma and liposarcoma type and histological grade. FINDINGS: 170 patients recruited between Oct 30, 2016, and Dec 23, 2020, were eligible in the discovery cohort and 89 patients recruited between Jan 18, 2012, and April 10, 2017, were eligible in the validation cohort. In the discovery cohort, the median age was 63 years (range 27-89), with 83 (49%) female and 87 (51%) male patients. In the validation cohort, median age was 59 years (range 33-77), with 46 (52%) female and 43 (48%) male patients. The highest performing model for the prediction of histological type had an area under the receiver operator curve (AUROC) of 0·928 on validation, based on a feature set of radiomics and approximate radiomic volume fraction. The highest performing model for the prediction of histological grade had an AUROC of 0·882 on validation, based on a radiomics feature set. INTERPRETATION: Our validated radiomics model can predict the histological type and grade of retroperitoneal sarcomas with excellent performance. This could have important implications for improving diagnosis and risk stratification in retroperitoneal sarcomas. FUNDING: Wellcome Trust, European Organisation for Research and Treatment of Cancer-Soft Tissue and Bone Sarcoma Group, the National Institutes for Health, and the National Institute for Health and Care Research Biomedical Research Centre at The Royal Marsden NHS Foundation Trust and The Institute of Cancer Research.

Prospective validation of a new imaging scorecard to assess leptomeningeal metastasis: A joint EORTC BTG and RANO effort.

BACKGROUND: Validation of the 2016 RANO MRI scorecard for leptomeningeal metastasis failed for multiple reasons. Accordingly, this joint EORTC Brain Tumor Group and RANO effort sought to prospectively validate a revised MRI scorecard for response assessment in leptomeningeal metastasis. METHODS: Coded paired cerebrospinal MRI of 20 patients with leptomeningeal metastases from solid cancers at baseline and follow-up after treatment and instructions for assessment were provided via the EORTC imaging platform. The Kappa coefficient was used to evaluate the interobserver pairwise agreement. RESULTS: Thirty-five raters participated, including 9 neuroradiologists, 17 neurologists, 4 radiation oncologists, 3 neurosurgeons, and 2 medical oncologists. Among single leptomeningeal metastases-related imaging findings at baseline, the best median concordance was noted for hydrocephalus (Kappa = 0.63), and the worst median concordance for spinal linear enhancing disease (Kappa = 0.46). The median concordance of raters for the overall response assessment was moderate (Kappa = 0.44). Notably, the interobserver agreement for the presence of parenchymal brain metastases at baseline was fair (Kappa = 0.29) and virtually absent for their response to treatment. 394 of 700 ratings (20 patients x 35 raters, 56%) were fully completed. In 308 of 394 fully completed ratings (78%), the overall response assessment perfectly matched the summary interpretation of the single ratings as proposed in the scorecard instructions. CONCLUSION: This study confirms the principle utility of the new scorecard, but also indicates the need for training of MRI assessment with a dedicated reviewer panel in clinical trials. Electronic case report forms with "blocking options" may be required to enforce completeness and quality of scoring.