Identifying Genetic Mutation Status in Patients with Colorectal Cancer Liver Metastases Using Radiomics-Based Machine-Learning Models.

For patients with colorectal cancer liver metastases (CRLM), the genetic mutation status is important in treatment selection and prognostication for survival outcomes. This study aims to investigate the relationship between radiomics imaging features and the genetic mutation status (KRAS mutation versus no mutation) in a large multicenter dataset of patients with CRLM and validate these findings in an external dataset. Patients with initially unresectable CRLM treated with systemic therapy of the randomized controlled CAIRO5 trial (NCT02162563) were included. All CRLM were semi-automatically segmented in pre-treatment CT scans and radiomics features were calculated from these segmentations. Additionally, data from the Netherlands Cancer Institute (NKI) were used for external validation. A total of 255 patients from the CAIRO5 trial were included. Random Forest, Gradient Boosting, Gradient Boosting + LightGBM, and Ensemble machine-learning classifiers showed AUC scores of 0.77 (95%CI 0.62-0.92), 0.77 (95%CI 0.64-0.90), 0.72 (95%CI 0.57-0.87), and 0.86 (95%CI 0.76-0.95) in the internal test set. Validation of the models on the external dataset with 129 patients resulted in AUC scores of 0.47-0.56. Machine-learning models incorporating CT imaging features could identify the genetic mutation status in patients with CRLM with a good accuracy in the internal test set. However, in the external validation set, the models performed poorly. External validation of machine-learning models is crucial for the assessment of clinical applicability and should be mandatory in all future studies in the field of radiomics.

Evaluation of a streamlined sentinel lymph-node imaging protocol in early-stage oral cancer.

OBJECTIVE: Sentinel lymph-node (SLN) mapping for early-stage oral squamous cell carcinoma (OSCC) is comprehensive and consequently time-consuming and costly. This study evaluated the clinical value of several SLN imaging components and analyzed the accuracy for SLN identification using a streamlined SLN imaging protocol in early-stage OSCC. MATERIALS AND METHODS: This retrospective within-patient evaluation study compared both number and localization of identified SLNs between the conventional SLN imaging protocol and a streamlined imaging protocol (dynamic lymphoscintigraphy (LSG) for 10 min directly post-injection and SPECT-CT at ~ 2 h post-injection). LSG and SPECT-CT images of 77 early-stage OSCC patients, scheduled for SLN biopsy, were evaluated by three observers. Identified SLNs using either protocol were related to histopathological assessment of harvested SLNs, complementary neck dissection specimens and follow-up status. RESULTS: A total of 200 SLNs were identified using the streamlined protocol, and 12 additional SLNs (n = 212) were identified with the conventional protocol in 10 patients. Of those, 9/12 were identified on early static LSG and 3/12 on late static LSG. None of the additionally identified SLNs contained metastases; none of those in whom additional SLNs were identified developed regional recurrence during follow-up. Only inferior alveolar process carcinoma showed a higher rate of additionally identified SLNs with the conventional protocol (p = 0.006). CONCLUSION: Early dynamic LSG can be reduced to 10 min. Late static LSG may be omitted, except in those with a history of oncological neck treatment or with OSCC featuring slow lymphatic drainage. Early static LSG appeared to be contributory in most OSCC subsites.