Use of machine learning algorithms to read thyroid nodule ultrasounds helpful, particularly for less experienced radiologists.
Machine learning algorithms may help less experienced radiologists differentiate malignant and benign thyroid nodules through ultrasound, according to a study published in the American Journal of Roentgenology.
Researchers from China sought to construct classifier models using machine learning algorithms and to evaluate their diagnostic performance for differentiating malignant from benign thyroid nodules.
Two radiologists respectively reviewed ultrasound images of 970 histopathologically proven thyroid nodules in 970 patients and their findings were compared with machine-building algorithms: the Naïve Bayes classifier, the support vector machine, and the radial basis neural function network. The nodules were graded according to a five-tier sonographic scoring system. The performance of the machine learning algorithms and radiologists were compared using ROC curve analysis. A total of 507 nodules (52.3%) were malignant, including 487 papillary thyroid carcinomas, 12 follicular thyroid carcinomas, four medullary thyroid carcinomas, three well-differentiated carcinomas, and one clear-cell carcinoma.
The results showed the experienced radiologist (17 years’ experience in reading thyroid ultrasounds) achieved the highest predictive accuracy of 88.7% with a specificity of 85.3%, whereas the radial basis function (RBF)–neural network (NN) achieved the highest sensitivity of 92.3%.
However, the algorithms were more accurate than the inexperienced radiologist, who only had three years’ similar experience:
“The machine learning algorithms underperformed with respect to the experienced radiologist's readings used to construct them,” the researchers concluded. “[A]nd the RBF-NN outperformed the other machine learning algorithm models.”
The Reading Room: Racial and Ethnic Minorities, Cancer Screenings, and COVID-19
November 3rd 2020In this podcast episode, Dr. Shalom Kalnicki, from Montefiore and Albert Einstein College of Medicine, discusses the disparities minority patients face with cancer screenings and what can be done to increase access during the pandemic.
Study Finds Transvaginal Ultrasound Unreliable for Detecting Endometrial Cancer in Black Patients
July 3rd 2024Utilizing a threshold of less than 5 mm of ultrasound-measured endometrial thickness, the authors of a new study noted an 11.4 percent false-negative probability for endometrial cancer in Black patients.
New Study Shows Non-Radiologists Interpreting 28 Percent of Imaging for Medicare Patients
June 28th 2024While radiologists interpreted approximately 99 percent of all non-cardiac CT, MRI and nuclear medicine studies in hospital and emergency department settings for Medicare beneficiaries, new research shows significantly less radiologist review of cardiac imaging and office-based imaging.
FDA Clears Pocket-Sized ECG System and AI Technology for Detection of Cardiac Conditions
June 27th 2024Using a reduced leadset and deep neural network algorithms trained on more than 175 million electrocardiograms, the KAI 12L technology reportedly detects up to 35 cardiac determinations, including acute myocardial infarction.