Detection of Liver Lesions in Colorectal Cancer Patients Using 18F-FDG PET/CT Dual-Time-Point Scan Imaging.

OBJECTIVE: The aim of this study was to evaluate the diagnostic performance of dual-time-point fluorine-18-fluorodeoxyglucose positron emission computed tomography/computed tomography (18F-FDG PET/CT) compared to conventional early imaging for detecting colorectal liver metastases (CRLM) in colorectal cancer (CRC) patients. METHODS: One hundred twenty-four consecutive CRC patients underwent dual-time-point imaging scans on a retrospective basis. Histopathological confirmation and/or clinical follow-up were accepted as the gold standard. Standard uptake values (SUV), signal-to-noise ratio (SNR), retention index (RI), tumor-to-normal liver ratio (TNR), and lesion sizes were measured for early and delayed PET scans. The diagnostic performance of early and delayed images was calculated on a per-patient basis and compared using McNemar's test. RESULTS: Among the 124 patients, 57 (46%) had CRLM, 6 (4.8%) had benign lesions, and 61 (49.2%) had no concerning lesions detected. Smaller CRLM lesions (<5 cm3) showed significantly higher uptake in the delayed scans relative to early imaging (p < 0.001). The SUV and TNR increased significantly in delayed imaging of all metastatic lesions (p < 0.001). The retention index of all CRLM was high (40.8%), especially for small lesions (54.8%). A total of 177 lesions in delayed images and 124 in standard early images were identified. In a per-patient analysis, delayed imaging had significantly higher sensitivity (100% vs. 87.7%) and specificity (91.0% vs. 94.0%) compared to early imaging (p-value = 0.04). CONCLUSIONS: The detection of liver lesions using dual-time-point PET/CT scan improves the sensitivity and specificity for the detection of colorectal liver metastasis.

Deep Learning Algorithms for Diagnosis of Lung Cancer: A Systematic Review and Meta-Analysis.

We conducted a systematic review and meta-analysis of the diagnostic performance of current deep learning algorithms for the diagnosis of lung cancer. We searched major databases up to June 2022 to include studies that used artificial intelligence to diagnose lung cancer, using the histopathological analysis of true positive cases as a reference. The quality of the included studies was assessed independently by two authors based on the revised Quality Assessment of Diagnostic Accuracy Studies. Six studies were included in the analysis. The pooled sensitivity and specificity were 0.93 (95% CI 0.85−0.98) and 0.68 (95% CI 0.49−0.84), respectively. Despite the significantly high heterogeneity for sensitivity (I2 = 94%, p < 0.01) and specificity (I2 = 99%, p < 0.01), most of it was attributed to the threshold effect. The pooled SROC curve with a bivariate approach yielded an area under the curve (AUC) of 0.90 (95% CI 0.86 to 0.92). The DOR for the studies was 26.7 (95% CI 19.7−36.2) and heterogeneity was 3% (p = 0.40). In this systematic review and meta-analysis, we found that when using the summary point from the SROC, the pooled sensitivity and specificity of DL algorithms for the diagnosis of lung cancer were 93% and 68%, respectively.