Novel T2 Mapping for Evaluating Cervical Cancer Features by Providing Quantitative T2 Maps and Synthetic Morphologic Images: A Preliminary Study
Li S, Liu J, Zhang F, et al.
Novel T2 Mapping for Evaluating Cervical Cancer Features by Providing Quantitative T2 Maps and Synthetic Morphologic Images: A Preliminary Study [published online ahead of print, 2020 Aug 14].
J Magn Reson Imaging. 2020;10.1002/jmri.27297. doi:10.1002/jmri.27297
https://doi.org/10.1002/jmri.27297
This retrospective study by Li et al evaluated the usefulness of T2 values in the classification, grading, and predicting the evidence of lymphovascular space invasion (LVSI) in patients with cervical cancer. The group evaluated 63 patients with histopathologically confirmed cervical cancer with T2W turbo spin echo, diffusion weighted echo-planar imaging, and accelerated T2 mapping sequence. The T2 mapping sequence provided both quantitative T2 maps as well as synthetic T2W images. The synthetic T2W images were generated from the values of the quantitative T2 maps and were compared to conventional T2W images. The T2 values between well/moderately differentiated tumors compared to poorly differentiated tumors were significantly different (92.8 vs 83.8 msec, p < 0.05). Additionally, tumors with LVSI also had lower T2 values (82.2 vs 93.9 msec, p < 0.05). ADC values also varied significantly between well/moderately differentiated and poorly differentiated tumor; however, was not useful for the evaluation of LVSI. The group suggests that the ability of accelerated T2 mapping to produce T2 maps and synthetic T2W images and the ability to predict LVSI makes it a useful tool for the evaluation and staging of cervical cancer.
PET/MRI in Cervical Cancer: Associations Between Imaging Biomarkers and Tumor Stage, Disease Progression, and Overall Survival
Shih IL, Yen RF, Chen CA, et al.
PET/MRI in Cervical Cancer: Associations Between Imaging Biomarkers and Tumor Stage, Disease Progression, and Overall Survival [published online ahead of print, 2020 Aug 14].
J Magn Reson Imaging. 2020;e27311. doi:10.1002/jmri.27311
https://doi.org/10.1002/jmri.27311
Shih et al conducted this prospective study on 54 patients with newly diagnosed cervical cancer to evaluate the associations between PET/MRI imaging biomarkers and tumor stage, progression free survival, and overall survival. A PET/MRI was obtained on 3T MRI scanner with diffusion weighted imaging at b values of 50 and 1000. Two radiologists measured ADCmin as well as the mean and max SUV, metabolic tumor volume (MTV), and total lesion glycolysis. Advanced tumors (T ≥ 1b2, M1, stage ≥ IB3) and tumors with nodal disease (N1) had lower ADCmin and high MTV, MTV/ADCmin. MaxSUV was a predictor of progression free survival, while ADCmin predicted overall survival. The group concluded that PET/MRI imaging biomarkers are associated with tumor stage and survival.
Machine Learning in Radiomic Renal Mass Characterization: Fundamentals, Applications, Challenges, and Future Directions
Kocak B, Kus EA, Yardimci AH, Bektas CT, Kilickesmez O.
Machine Learning in Radiomic Renal Mass Characterization: Fundamentals, Applications, Challenges, and Future Directions [published online ahead of print, 2020 Aug 12].
AJR Am J Roentgenol. 2020;1-9. doi:10.2214/AJR.19.22608
https://doi.org/10.2214/AJR.19.22608
This is a review article that provides an overview of the use of machine learning (ML) and deep learning approaches in the characterization of renal masses. The review covers fundamental concepts, current applications, and challenges to the implementation of machine learning in the characterization of renal masses. The article describes that the availability of many open-source, free, and easy-to-use toolboxes have created a low barrier of entry of ML into radiology with current uses in renal mass characterization to include determination of malignant from benign lesions, tumor aggressiveness, tumor subtyping, and radiogenomics. The authors suggest that larger datasets and new algorithms will lead to limitless possibilities to help move the field forward.
Multicenter Multireader Evaluation of an Artificial Intelligence–Based Attention Mapping System for the Detection of Prostate Cancer With Multiparametric MRI
Mehralivand S, Harmon SA, Shih JH, et al.
Multicenter Multireader Evaluation of an Artificial Intelligence-Based Attention Mapping System for the Detection of Prostate Cancer With Multiparametric MRI [published online ahead of print, 2020 Aug 5].
AJR Am J Roentgenol. 2020;1-10. doi:10.2214/AJR.19.22573
https://doi.org/10.2214/AJR.19.22573
Mehralivand et al performed a study evaluating 152 patients with 274 proven prostate cancer lesions from 5 institutions. 9 different readers evaluated the mpMRI studies using PIRADsv2 with and without an AI based detection system. When the readers used the AI based system, they were only able to accept or reject the lesions that were within the AI generated attention boxes. The overall sensitivity for the detection of lesions was higher for the AI based system (57.4% vs 53.6% p = 0.073) with the AUC of 77.5% for AI and 74.9% for mpMRI alone. There was moderate inter-reader agreement between AI and mpMRI alone with greater sensitivity for the lesion detection in the transition zone for the AI based system. Overall the group suggested that the AI based system showed significant improvement in the detection of transition zone lesions.
Utilization and Yield of CT Urography: Are the American Urological Association Guidelines for Imaging Patients With Asymptomatic Microscopic Hematuria Being Followed?
Skaggs, A. W., Loehfelm, T. W., Fananapazir, G., Dall’era, M., & Corwin, M. T. (2020).
Utilization and Yield of CT Urography: Are the American Urological Association Guidelines for Imaging Patients With Asymptomatic Microscopic Hematuria Being Followed?
American Journal of Roentgenology. doi:10.2214/ajr.20.22998
http://doi.org/10.2214/ajr.20.22998
Skagg et al performed a retrospective review of 419 patients with asymptomatic microscopic hematuria. Patients with benign etiology of hematuria or prior urologic malignancy were excluded. Of the 419 patients, 200 were selected to determine appropriateness, which was defined as urinalysis with > 3 rbcs per high power field. Approximately 29% of the randomly sampled 200 patients did not meet AUA criteria for radiologic evaluation. 1 solid renal mass and 1 possible bladder mass without any evidence of upper tract urothelial neoplasm were identified.
Vesicle Imaging and Data Reporting System (VI-RADS): Multi-institutional multi-reader diagnostic accuracy and inter-observer agreement study
Ueno Y, Tamada T, Takeuchi M, et al.
Vesicle Imaging and Data Reporting System (VI-RADS): Multi-institutional multi-reader diagnostic accuracy and inter-observer agreement study [published online ahead of print, 2020 Jul 29].
AJR Am J Roentgenol. 2020;10.2214/AJR.20.23604. doi:10.2214/AJR.20.23604
https://doi.org/10.2214/AJR.20.23604
This group conducted a retrospective study of 91 patients that aimed to evaluate the interobserver agreement and diagnostic performance of Vesical-Imaging Reporting and Data System (VI-RADS) in predicting the presence of muscle-invasive bladder cancer (MIBC). VI-RADS is a 5-point scoring system with the aim of establishing the likelihood of MIBC with a score of 4/5 suggesting that muscle invasion is likely. The presence of muscle invasion ultimately requires patients to undergo radical cystectomy. In this study the muscle invasion status of the tumor was determined using histologic slices from TURBT or cystectomy, and this was compared to VI-RADS scores that were assigned by 7 radiologists of varying experience. There was moderate to substantial agreement amongst experienced readers and between experienced and inexperienced readers. The sensitivity and specificity of obtaining a score of 4 or greater is 74.1% and 94.1% for experienced readers and 63.9 % and 86.4% for inexperienced readers.
Prevalence of prostate cancer in PI-RADS version 2.1 transition zone ‘atypical nodules’ upgraded by abnormal diffusion weighted imaging: correlation with MRI-directed TRUS-guided targeted biopsy.
Lim CS, Abreu-Gomez J, Carrion I, Schieda N.
Prevalence of prostate cancer in PI-RADS version 2.1 transition zone ‘atypical nodules’ upgraded by abnormal diffusion weighted imaging: correlation with MRI-directed TRUS-guided targeted biopsy [published online ahead of print, 2020 Jul 29].
AJR Am J Roentgenol. 2020;10.2214/AJR.20.23932. doi:10.2214/AJR.20.23932
http://doi.org/10.2214/AJR.20.23932
Lim et al did a retrospective analysis in 104 consecutive men who underwent MR guided biopsy of transition zone atypical nodules that were upgraded using DWI. There were 95 PI-RADs category 3 lesions, 40 of which were upgraded based on DWI. There was slight to substantial agreement between radiologists when categorizing each of the lesions. Of the 40 upgraded nodules only 27.5% had prostate cancer and 7.5% had clinically significant prostate cancer compared to 43.6% and 20.0% nodules that were diagnosed as PI-RADs category 3 using T2W imaging. The group suggests that prevalence of PCa in DWI upgraded TZ atypical nodules compared favorably to conventionalT2 weighted imaging, which validates the use of the DWI upgrade rules introduced in PIRADSv2.1.
Diagnostic accuracy of 68Ga-PSMA PET/CT for initial detection in patients with suspected prostate cancer: A systematic review and meta-analysis
Satapathy, S., Singh, H., Kumar, R., & Mittal, B. R. (2020).
Diagnostic accuracy of 68Ga-PSMA PET/CT for initial detection in patients with suspected prostate cancer: A systematic review and meta-analysis.
American Journal of Roentgenology. doi:10.2214/ajr.20.23912
https://doi.org/10.2214/ajr.20.23912
The authors performed a systematic review evaluating the performance of GA-PSMA PET-CT . The group used 7 articles comprising 389 patients with a pooled sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of 0.97, 0.66, 2.86, and 0.05 respectively. They showed that GA-PSMA PET CT had excellent sensitivity and high negative likelihood ratio in the initial diagnosis of prostate cancer.
References
