ABSTRACT
OBJECTIVES: To validate a deep learning (DL) algorithm for measurement of skeletal muscular index (SMI) and prediction of overall survival in oncology populations. METHODS: A retrospective single-center observational study included patients with metastatic renal cell carcinoma between 2007 and 2019. A set of 37 patients was used for technical validation of the algorithm, comparing manual vs DL-based evaluations. Segmentations were compared using mean Dice similarity coefficient (DSC), SMI using concordance correlation coefficient (CCC) and Bland-Altman plots. Overall survivals (OS) were compared using log-rank (Kaplan-Meier) and Mann-Whitney tests. Generalizability of the prognostic value was tested in an independent validation population (N = 87). RESULTS: Differences between two manual segmentations (DSC = 0.91, CCC = 0.98 for areas) or manual vs. automated segmentation (DSC = 0.90, CCC = 0.98 for areas, CCC = 0.97 for SMI) had the same order of magnitude. Bland-Altman plots showed a mean difference of -3.33 cm2 [95%CI: -15.98, 9.1] between two manual segmentations, and -3.28 cm2 [95% CI: -14.77, 8.21] for manual vs. automated segmentations. With each method, 20/37 (56%) patients were classified as sarcopenic. Sarcopenic vs. non-sarcopenic groups had statistically different survival curves with median OS of 6.0 vs. 12.5 (p = 0.008) and 6.0 vs. 13.9 (p = 0.014) months respectively for manual and DL methods. In the independent validation population, sarcopenic patients according to DL had a lower OS (10.7 vs. 17.3 months, p = 0.033). CONCLUSION: A DL algorithm allowed accurate estimation of SMI compared to manual reference standard. The DL-calculated SMI demonstrated a prognostic value in terms of OS. KEY POINTS: ⢠A deep learning algorithm allows accurate estimation of skeletal muscle index compared to a manual reference standard with a concordance correlation coefficient of 0.97. ⢠Sarcopenic patients according to SMI thresholds after segmentation by the deep learning algorithm had statistically significantly lower overall survival compared to non-sarcopenic patients.
Subject(s)
Carcinoma, Renal Cell , Deep Learning , Kidney Neoplasms , Sarcopenia , Algorithms , Carcinoma, Renal Cell/complications , Carcinoma, Renal Cell/diagnostic imaging , Carcinoma, Renal Cell/pathology , Humans , Kidney Neoplasms/complications , Kidney Neoplasms/pathology , Muscle, Skeletal/diagnostic imaging , Muscle, Skeletal/pathology , Retrospective Studies , Sarcopenia/complications , Sarcopenia/diagnostic imagingABSTRACT
Purpose To assess the performance of the ITK-SNAP software for fluorodeoxyglucose (FDG) positron emission tomography (PET) segmentation of complex-shaped lung tumors compared with an optimized, expert-based manual reference standard. Materials and Methods Seventy-six FDG PET images of thoracic lesions were retrospectively segmented by using ITK-SNAP software. Each tumor was manually segmented by six raters to generate an optimized reference standard by using the simultaneous truth and performance level estimate algorithm. Four raters segmented 76 FDG PET images of lung tumors twice by using ITK-SNAP active contour algorithm. Accuracy of ITK-SNAP procedure was assessed by using Dice coefficient and Hausdorff metric. Interrater and intrarater reliability were estimated by using intraclass correlation coefficients of output volumes. Finally, the ITK-SNAP procedure was compared with currently recommended PET tumor delineation methods on the basis of thresholding at 41% volume of interest (VOI; VOI41) and 50% VOI (VOI50) of the tumor's maximal metabolism intensity. Results Accuracy estimates for the ITK-SNAP procedure indicated a Dice coefficient of 0.83 (95% confidence interval: 0.77, 0.89) and a Hausdorff distance of 12.6 mm (95% confidence interval: 9.82, 15.32). Interrater reliability was an intraclass correlation coefficient of 0.94 (95% confidence interval: 0.91, 0.96). The intrarater reliabilities were intraclass correlation coefficients above 0.97. Finally, VOI41 and VOI50 accuracy metrics were as follows: Dice coefficient, 0.48 (95% confidence interval: 0.44, 0.51) and 0.34 (95% confidence interval: 0.30, 0.38), respectively, and Hausdorff distance, 25.6 mm (95% confidence interval: 21.7, 31.4) and 31.3 mm (95% confidence interval: 26.8, 38.4), respectively. Conclusion ITK-SNAP is accurate and reliable for active-contour-based segmentation of heterogeneous thoracic PET tumors. ITK-SNAP surpassed the recommended PET methods compared with ground truth manual segmentation.
Subject(s)
Fluorodeoxyglucose F18 , Image Processing, Computer-Assisted/methods , Lung Neoplasms/diagnostic imaging , Positron-Emission Tomography/methods , Radiopharmaceuticals , Algorithms , Female , Humans , Lung/diagnostic imaging , Male , Reproducibility of Results , Retrospective Studies , SoftwareABSTRACT
Common mesentery is an abnormal rotation of the primary umbilical loop characterized by inverted positioning of the mesenteric vessels; the mesenteric vein is displaced to the left of the artery. The inversion can be complete or incomplete. If it is incomplete, the mesenteric root is very short, with an empty right iliac fossa and the caecum in high median or subhepatic position. If it is complete, the entire small intestine is on the right, the entire large intestine is on the left; there is no third duodenum, and the second duodenum is anastomosed in the jejunum to the right of the superior mesenteric vessels. Cecal volvulus is a rarely encountered cause of acute intestinal occlusion and should be considered as a surgical emergency. There exist two main types of volvulus: by twisting of the large intestine around its axis, which remains in place; or by tilt and to rotation of the colon, which changes position.
Subject(s)
Cecal Diseases , Intestinal Volvulus , Mesentery , Intestinal Volvulus/surgery , Intestinal Volvulus/diagnostic imaging , Intestinal Volvulus/complications , Humans , Cecal Diseases/surgery , Cecal Diseases/diagnostic imaging , Mesentery/surgery , Male , Female , Tomography, X-Ray ComputedABSTRACT
There are no current recommendations on which machine learning (ML) algorithms should be used in radiomics. The objective was to compare performances of ML algorithms in radiomics when applied to different clinical questions to determine whether some strategies could give the best and most stable performances regardless of datasets. This study compares the performances of nine feature selection algorithms combined with fourteen binary classification algorithms on ten datasets. These datasets included radiomics features and clinical diagnosis for binary clinical classifications including COVID-19 pneumonia or sarcopenia on CT, head and neck, orbital or uterine lesions on MRI. For each dataset, a train-test split was created. Each of the 126 (9 × 14) combinations of feature selection algorithms and classification algorithms was trained and tuned using a ten-fold cross validation, then AUC was computed. This procedure was repeated three times per dataset. Best overall performances were obtained with JMI and JMIM as feature selection algorithms and random forest and linear regression models as classification algorithms. The choice of the classification algorithm was the factor explaining most of the performance variation (10% of total variance). The choice of the feature selection algorithm explained only 2% of variation, while the train-test split explained 9%.