Radiomics signature from [18F]FDG PET images for prognosis predication of primary gastrointestinal diffuse large B cell lymphoma

Prognostic value of 18F-FDG PET radiomics and sarcopenia in patients with oral squamous cell carcinoma

BACKGROUND

Oral cancer is one of the most common malignancies in the head and neck region. Approximately 90% of oral cancers are oral squamous cell carcinomas (OSCC). 18F-FDG PET/CT has been used in OSCC patients for its high value in detecting metastatic lymph nodes and distant metastases. PET radiomics and sarcopenia can be measured on the PET and CT components of 18F-FDG PET/CT.

PURPOSE

This study aimed to investigate the prognostic value of radiomics and sarcopenia measured on the PET and CT components of pre-operation 18F-FDG PET/CT in OSCC.

METHODS

A total of 116 patients eventually enrolled in our study were randomly divided into two cohorts: training cohort (n = 58) and validation cohort (n = 58). The Cox model combined with the least absolute shrinkage and selection operator (LASSO) algorithm was applied to construct the radiomics score (Rad_score). The third lumber skeletal muscle index (L3 SMI) was calculated to identify sarcopenia. Univariate and multivariate Cox regression analyses were performed to identify the independent prognostic factors. Based on the clinical factors, the clinical model was constructed, and the combined model was developed through the combination of the clinical model and Rad_score. C index, time-dependent C-index curves, receiver operating characteristic (ROC) curve, calibration curves, and decision curve analysis were used to evaluate the performance of prediction models.

RESULTS

Three radiomics features constitute the Rad_score for overall survival (OS) and progression-free survival (PFS), respectively. Multivariate Cox regression analysis revealed that Rad_score was an independent prognostic factor, whereas sarcopenia was not. The combined models showed satisfactory performance in both the training cohort (C-index: OS:0.836, PFS:0.776) and the validation cohort (C-index: OS:0.744, PFS:0.712). The combined models were visualized as nomograms. Nomogram scores can realize the risk stratification of OSCC patients. Lower nomogram score is significantly related to the poorer OS (training cohort: p < 0.0001, validation cohort: p < 0.0001, overall cohort: p < 0.0001) and PFS (training cohort: p < 0.0001, validation cohort: p = 0.00017, overall cohort: p < 0.0001).

CONCLUSIONS

Rad_score, but not sarcopenia, was an independent prognostic factor for patients with OSCC. The nomograms had a satisfactory performance, which might be helpful for OSCC patients and clinicians in personalized prognostic prediction and treatment decision-making.

Clinical application of 18F-fluorodeoxyglucose positron emission tomography/computed tomography radiomics-based machine learning analyses in the field of oncology

Machine learning (ML) analyses using 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)/computed tomography (CT) radiomics features have been applied in the field of oncology. The current review aimed to summarize the current clinical articles about 18F-FDG PET/CT radiomics-based ML analyses to solve issues in classifying or constructing prediction models for several types of tumors. In these studies, lung and mediastinal tumors were the most commonly evaluated lesions, followed by lymphatic, abdominal, head and neck, breast, gynecological, and other types of tumors. Previous studies have commonly shown that 18F-FDG PET radiomics-based ML analysis has good performance in differentiating benign from malignant tumors, predicting tumor characteristics and stage, therapeutic response, and prognosis by examining significant differences in the area under the receiver operating characteristic curves, accuracies, or concordance indices (> 0.70). However, these studies have reported several ML algorithms. Moreover, different ML models have been applied for the same purpose. Thus, various procedures were used in 18F-FDG PET/CT radiomics-based ML analysis in oncology, and 18F-FDG PET/CT radiomics-based ML models, which are easy and universally applied in clinical practice, would be expected to be established.

Advances in diagnosis, treatment and prognostic factors of gastrointestinal DLBCL

Gastrointestinal diffuse large B-cell lymphoma (GI-DLBCL) is an extremely aggressive form of B-cell non-Hodgkin lymphoma (BNHL) which has complex histological characteristics and manifests a high degree of heterogeneity in terms of clinical, morphological, immunological, and genetic features. GI-DLBCL mainly spreads by infiltrating neighboring lymph nodes, and common gastrointestinal complications (GICS) such as obstruction, perforation, or bleeding, frequently arise during the progression of the disease, posing significant challenges in both diagnosing and treating the condition. Meanwhile, the incidence of GI-DLBCL has been gradually increasing in recent years, and its strong invasiveness makes it prone to being misdiagnosed or completely missed. In clinical practice, over half of the patients diagnosed with the disease are in stage III or stage IV. What makes it worse is that certain patients may not exhibit a favorable response to chemotherapy. All these lead to intricacies in management of this disease. Unfortunately, there is currently no large prospective study or evidence-based medical evidence to provide clear guidance on treatment decisions for this specific type of lymphoma. Neither do physicians have a consensus regarding the optimal approach to address this condition. Recent studies have identified the presence of various prognostic factors that significantly impact survival in GI-DLBCL, which demonstrates the unique particularity of GI-DLBCL, and could help optimize the clinical decision.

Role of Artificial Intelligence in PET/CT Imaging for Management of Lymphoma

Our review shows that AI-based analysis of lymphoma whole-body FDG-PET/CT can inform all phases of clinical management including staging, prognostication, treatment planning, and treatment response evaluation. We highlight advancements in the role of neural networks for performing automated image segmentation to calculate PET-based imaging biomarkers such as the total metabolic tumor volume (TMTV). AI-based image segmentation methods are at levels where they can be semi-automatically implemented with minimal human inputs and nearing the level of a second-opinion radiologist. Advances in automated segmentation methods are particularly apparent in the discrimination of lymphomatous vs non-lymphomatous FDG-avid regions, which carries through to automated staging. Automated TMTV calculators, in addition to automated calculation of measures such as Dmax are informing robust models of progression-free survival which can then feed into improved treatment planning.