Search

For accurate diagnosis of interstitial lung disease (ILD), a consensus of radiologic, pathological, and clinical findings is vital. Management of ILD also requires thorough follow-up with computed tomography (CT) studies and lung function tests to assess disease progression, severity, and response to treatment. However, accurate classification of ILD subtypes can be challenging, especially for those not accustomed to reading chest CTs regularly. Dynamic models to predict patient survival rates based on longitudinal data are challenging to create due to disease complexity, variation, and irregular visit intervals. Here, we utilize RadImageNet pretrained models to diagnose five types of ILD with multimodal data and a transformer model to determine a patient's 3-year survival rate. When clinical history and associated CT scans are available, the proposed deep learning system can help clinicians diagnose and classify ILD patients and, importantly, dynamically predict disease progression and prognosis.

Magnetic resonance imaging (MRI) is a ubiquitous medical imaging technology with applications in disease diagnostics, intervention, and treatment planning. Accurate MRI segmentation is critical for diagnosing abnormalities, monitoring diseases, and deciding on a course of treatment. With the advent of advanced deep learning frameworks, fully automated and accurate MRI segmentation is advancing. Traditional supervised deep learning techniques have advanced tremendously, reaching clinical‐level accuracy in the field of segmentation. However, these algorithms still require a large amount of annotated data, which is oftentimes unavailable or impractical. One way to circumvent this issue is to utilize algorithms that exploit a limited amount of labeled data. This paper aims to review such state‐of‐the‐art algorithms that use a limited number of annotated samples. We explain the fundamental principles of self‐supervised learning, generative models, few‐shot learning, and semi‐supervised learning and summarize their applications in cardiac, abdomen, and brain MRI segmentation. Throughout this review, we highlight algorithms that can be employed based on the quantity of annotated data available. We also present a comprehensive list of notable publicly available MRI segmentation datasets. To conclude, we discuss possible future directions of the field—including emerging algorithms, such as contrastive language‐image pretraining, and potential combinations across the methods discussed—that can further increase the efficacy of image segmentation with limited labels. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI), particularly deep learning, has made enormous strides in medical imaging analysis. In the field of musculoskeletal radiology, deep-learning models are actively being developed for the identification and evaluation of bone fractures. These methods provide numerous benefits to radiologists such as increased diagnostic accuracy and efficiency while also achieving standalone performances comparable or superior to clinician readers. Various algorithms are already commercially available for integration into clinical workflows, with the potential to improve healthcare delivery and shape the future practice of radiology. In this systematic review, we explore the performance of current AI methods in the identification and evaluation of fractures, particularly those in the ankle, wrist, hip, and ribs. We also discuss current commercially available products for fracture detection and provide an overview of the current limitations of this technology and future directions of the field. [ABSTRACT FROM AUTHOR]

Introduction: The direct anterior approach (DAA) and the posterior approach (PA) are 2 common total hip arthroplasty (THA) exposures. This prospective study quantitatively compared changes in periarticular muscle volume after DAA and PA THA. Materials: 19 patients undergoing THA were recruited prospectively from the practices of 3 fellowship-trained hip surgeons. Each surgeon performed a single approach, DAA or PA. Enrolled patients underwent a preoperative MRI of the affected hip and two subsequent postoperative MRIs at around 6 weeks and 6 months after surgery. Clinical evaluations were done by Harris Hip Score at each follow-up interval. Results: MRIs or 10 DAA and 9 PA patients were analysed. Groups did not differ significantly with regard to BMI, age, or preoperative muscle volume. 1 DAA patient suffered a periprosthetic fracture and was excluded from the study. DAA hips showed significant atrophy in the obturator internus (−37.3%) muscle at early follow-up, with persistent atrophy of this muscle at the final follow-up. PA hips showed significant atrophy in the obturator internus (−46.8%) and externus (−16.0%), piriformis (−8.12%), and quadratus femoris muscles (−13.1%) at early follow-up, with persistent atrophy of these muscles at final follow-up. Loss of anterior capsular integrity was present at final follow-up in 2/10 DAA hips while loss of posterior capsular integrity was present in 5/9 PA hips. There was no difference in clinical outcomes. Discussion: This study demonstrates that DAA showed less persistent muscular atrophy than PA. Regardless of surgical approach, a muscle whose tendon is detached from its insertion is likely to demonstrate persistent atrophy 6 months following THA. Although the study was not powered to compare clinical outcomes, it should be noted that no significant difference in patient outcomes was observed. [ABSTRACT FROM AUTHOR]

Background: There is a paucity of clinical studies that compare the efficacy of autograft and allograft in osteochondral transplantation for treatment of osteochondral lesions of the talus (OLT). The purpose of the present study was to compare the clinical and radiographic outcomes following osteochondral transplantation with autograft or allograft for OLT.Methods: A retrospective analysis comparing patients treated with autograft or allograft for OLT was performed. Clinical outcomes were evaluated with use of the Foot and Ankle Outcome Score (FAOS) and the Short Form-12 (SF-12) score. Magnetic resonance imaging (MRI) was evaluated with use of the Magnetic Resonance Observation of Cartilage Repair Tissue (MOCART) score. The rates of cyst occurrence, graft degradation, graft failure, and revision surgeries were also evaluated.Results: Twenty-five nonrandomized patients with autograft and 16 with allograft were included, with a mean follow-up of 26 months in the autograft group and 22 months in the allograft group. There were no significant differences among all demographic variables between the autograft and allograft groups. The mean postoperative FAOS was significantly higher in the autograft group (81.9; 95% confidence interval [CI]: 78.6 to 85.2) than in the allograft group (70.1; 95% CI: 63.7 to 76.5; p = 0.006). Similarly, the mean postoperative SF-12 scores were significantly higher in the autograft group (74.7; 95% CI: 71.0 to 78.4) than in the allograft group (66.1; 95% CI: 61.2 to 71.0; p = 0.021). MOCART scores were significantly better in the autograft group (87.1) than in the allograft group (75.5; p = 0.005). The rate of chondral wear on MRI was higher in the allograft group (53%) than in the autograft group (4%; p < 0.001). Cyst formation in the graft itself was more likely to occur in the allograft group (47%) than in the autograft group (8%; p = 0.017). The rate of secondary procedures for the graft was higher in the allograft group (25%) than in the autograft group (0%; p = 0.009).Conclusions: In this small nonrandomized cohort study, the procedures performed with use of an autograft provided better clinical and MRI outcomes than the allograft procedures. The rate of chondral wear on MRI was higher with allograft than with autograft, and allograft-treated patients had a higher rate of clinical failure.Level Of Evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence. [ABSTRACT FROM AUTHOR]

Background: Subchondral bone marrow edema (BME) has been associated with articular cartilage loss, with the potential to be a negative prognostic indicator for clinical outcomes after microfracture. However, no single study has investigated the association between BME and clinical outcomes after microfracture for osteochondral lesions of the talus (OLTs) at midterm follow-up. Purpose: To clarify the association between postoperative subchondral BME and clinical outcomes in patients treated with microfracture for OLTs at both short-term and midterm follow-up using a grading system that classified the extent of BME of the talus. Study Design: Cohort study; Level of evidence, 3. Methods: Patients who underwent microfracture between 2008 and 2013 were assessed at 2- and 4-year postoperative follow-up. BME was evaluated using magnetic resonance imaging, and the presence of subchondral BME was determined with fat-suppressed T2-weighted sequences. Clinical outcomes were evaluated using the Foot and Ankle Outcome Score (FAOS). P < .05 was considered to be statistically significant. Results: Forty-three (83%) of 52 eligible patients were included. No significant differences were found in the FAOS between the BME and no BME groups at 2-year follow-up (83.1 ± 6.5 vs 88.6 ± 8.0, respectively; P = .109), but there was a significant difference at 4-year follow-up (77.5 ± 11.1 vs 84.7 ± 8.4, respectively; P = .041). A significant difference was found among BME grades at 4-year follow-up (grade 0: 84.7 ± 7.4, grade 1: 80.1 ± 10.5, grade 2: 74.0 ± 10.3, and grade 3: 67.5 ± 7.1; P = .035). A post hoc analysis showed significant differences between grades 0 and 2, 0 and 3, and 1 and 3 (P = .041, .037, and .048, respectively). In addition, at 4-year follow-up, a significant correlation was noted between the FAOS and BME grade (r = −0.453, P = .003) but not at 2-year follow-up (r = −0.212, P = .178). Seventy-four percent of patients still had subchondral BME at 4-year follow-up after microfracture for OLTs. Conclusion: Patients with subchondral BME at midterm follow-up after microfracture for OLTs had worse clinical outcomes than those without subchondral BME. In addition, the degree of subchondral BME at midterm follow-up was correlated with clinical outcomes. However, at short-term follow-up, there were no significant differences in clinical outcomes based on both the presence and degree of BME, and no correlation was found between clinical outcomes and the degree of BME. The current study suggests that BME at short-term follow-up is a normal physiological reaction. However, BME at midterm follow-up after microfracture for OLTs may be pathological and is associated with poorer clinical outcomes. [ABSTRACT FROM AUTHOR]

Background: Uncontained-type osteochondral lesions of the talus (OLTs) have been shown to have inferior clinical outcomes after treatment with bone marrow stimulation. While autologous osteochondral transplantation (AOT) is indicated for larger lesions, no study has reported on the prognostic significance of the containment of OLTs treated with the AOT procedure. Purpose: To clarify the effect of the containment of OLTs on clinical and radiological outcomes in patients who underwent AOT for OLTs. Study Design: Case control study; Level of evidence, 3. Methods: A retrospective cohort study comparing patients with contained-type and uncontained-type OLTs was undertaken to include all patients who underwent AOT for the treatment of OLTs between 2006 and 2014. Analyses were performed by grouping the patients according to the containment type. Clinical outcomes were evaluated using the Foot and Ankle Outcome Score (FAOS) and the 12-Item Short Form Health Survey (SF-12) preoperatively and at final follow-up. Magnetic resonance imaging (MRI) at 2 years’ follow-up was evaluated with the modified magnetic resonance observation of cartilage repair tissue (MOCART) score. Multivariate regression models were used to evaluate factors affecting postoperative FAOS, SF-12, and MOCART scores. Results: Ninety-four patients were included: 31 patients with a contained-type OLT and 63 patients with an uncontained-type OLT. The median patient age was 34 years (interquartile range [IQR], 28-48 years) in the contained-type group and 36 years (IQR, 27-46 years) in the uncontained-type group. The median follow-up time was 45 months (IQR, 38-63 months) in the contained-type group and 52 months (IQR, 40-66 months) in the uncontained-type group. The median FAOS and SF-12 scores improved significantly after surgery in both contained-type and uncontained-type lesions (P < .001). The median postoperative FAOS score of patients with contained-type OLTs was higher than that of patients with uncontained-type OLTs (91.7 vs 85.0, respectively; P = .009), but no significant differences were found between the contained-type and uncontained-type groups for postoperative SF-12 and MOCART scores. The multivariate regression models showed that patients with contained-type OLTs had an approximately 10-point better score on the FAOS compared with patients with uncontained-type OLTs (P = .006). There was a nonsignificant trend for the rate of cystic occurrence in uncontained-type OLTs to be higher than that of contained-type OLTs (55.6% vs 38.7%, respectively; P = .125). Conclusion: Patients with contained-type OLTs experienced better clinical outcomes than those with uncontained-type OLTs after AOT for the treatment of OLTs. However, the AOT procedure still provided good clinical and MRI outcomes in both contained-type and uncontained-type OLTs at midterm follow-up. [ABSTRACT FROM AUTHOR]

Background: Autologous osteochondral transplantation (AOT) has an inherent risk of donor site morbidity (DSM). The reported rates of DSM vary from 0% to 50%, with few studies reporting clinical or imaging outcomes at the donor site as a primary outcome and even fewer report these outcomes when a biosynthetic plug backfill is employed. Although TruFit (Smith & Nephew, Andover, MA) plugs have been removed from the market for regulatory purposes, biphasic plugs (including TruFit plugs) have been used for several years and the evaluation of these is therefore pertinent. Methods: Thirty-nine patients who underwent forty AOT procedures of the talus, with the donor graft being taken from the ipsilateral knee, were included. Postoperative magnetic resonance imaging (MRI) was used to assess the donor site graded with magnetic resonance observation of cartilage repair tissue (MOCART) scoring. Lysholm scores were collected preoperatively, at the time of magnetic resonance imaging (MRI), and again at 24 months and at final follow-up to assess clinical outcomes. Statistical analysis was performed to establish if there was any correlation between MRI assessment of the donor site and clinical outcomes. The mean patient age was 36.2 ± 15.7 years with a mean follow-up of 41.8 ± 16.7 months. Results: All patient donor site defects were filled with OBI TruFit biphasic plugs. DSM was encountered in 12.5% of the patient cohort at 24 months, and in these patients, the Lysholm score was a mean 87.2 ± 5.0. At final follow-up, DSM was reduced to 5%. Lysholm scores for the entire cohort were 98.4 ± 4.6 and 99.4 ± 3.1 at 24 months and final follow-up, respectively. MRI of the donor sites were taken at an average of 18.1 ± 13.5 (range, 3-48) months postoperatively and the mean MOCART score was 60.0 ± 13.5. No correlation was found between the MOCART score and Lysholm outcomes at the donor knee (P = .43, r = 0.13). Conclusion: Low incidence of DSM and good functional outcomes were achieved with AOT. Additionally, MRI findings did not predict clinical outcomes in our study. Level of Evidence: Level IV, retrospective case series. [ABSTRACT FROM AUTHOR]

Background: Autologous osteochondral transplantation (AOT) is used to treat osteochondral lesions (OCLs) of the talus, typically reserved for lesions greater than 150 mm2. Few studies exist examining the functional and magnetic resonance imaging (MRI) outcomes following this procedure. The purpose of this study was to investigate functional and MRI outcomes, including quantitative T2 mapping following AOT. Methods: Eighty-five consecutive patients who underwent AOT were identified. Functional outcomes were assessed pre- and postoperatively using the Foot and Ankle Outcome Score (FAOS). The Magnetic Resonance Observation of Cartilage Repair Tissue (MOCART) score was used to assess cartilage incorporation. Quantitative T2 MRI relaxation time of graft tissue and adjacent normal cartilage values were recorded in a subset of 61 patients. The mean clinical follow-up was 47.2 months, with mean MRI follow-up of 24.8 months. Results: Mean FAOS improved pre- to postoperatively from 50 to 81 (P < .001). The mean MOCART score was 85.8. Lesion size was negatively correlated with MOCART score (r = −0.36, P = .004). Superficial T2 values in graft tissue were higher than control tissue (42.0 vs 35.8, P < .001). Deep T2 values in graft tissue were similar to the control values (30.9 vs 30.0, P = .305). Functional outcomes were similar in patients irrespective of whether they had previous microfracture or concomitant procedures. Conclusion: AOT was an effective treatment for large OCLs of the talus in the current study. MOCART scoring indicated good structural integrity of the graft. Quantitative T2 mapping suggests that graft tissue may not always mirror native hyaline cartilage. The long-term implications of this are not yet known. Level of Evidence: Level IV, case series. [ABSTRACT FROM AUTHOR]

Background: Autologous osteochondral transplantation (AOT) is a primary treatment strategy for large or cystic osteochondral lesions of the talus (OLT) or a secondary replacement strategy after failed bone marrow stimulation. The technique requires perpendicular access to the talar dome, which is often difficult to obtain for posterior or lateral lesions. Traditional methods to access these areas have required disruption of the syndesmotic complex with concern over osteotomy reduction, malalignment, and ligament disruption. An alternate to these traditional methods of access is an anterolateral tibial osteotomy. The purpose of this study was to report functional and magnetic resonance imaging (MRI) outcomes in a series of patients that underwent AOT for treatment of an OLT via an anterolateral tibial osteotomy. Methods: Records of patients that underwent an anterolateral tibial osteotomy for AOT were retrospectively reviewed. Pre- and postoperative Foot and Ankle Outcome Scores (FAOS) and demographic data were recorded. Magnetic resonance observation of cartilage repair tissue (MOCART) was used to assess morphologic state of tibial cartilage at the repair site of the osteotomy. Quantitative T2 mapping MRI was analyzed in the superficial and deep cartilage layers of the repair site of the osteotomy and in adjacent normal cartilage to serve as control tissue. Seventeen patients with a mean age of 36.9 (range, 17-76) years underwent anterolateral tibial osteotomy with a mean follow-up of 64 (range, 29 to 108) months. MOCART data were available in 9 of 17 patients, and quantitative T2 mapping was available in 6 patients. Results: FAOS significantly improved from an average 39.2 (range, 14 to 66) out of 100 points preoperatively to 81.2 (range, 19 to 98) postoperatively (P < .01). The average MOCART score was 73.9 out of 100 points (range, 40 to 100). Quantitative T2 analysis demonstrated relaxation times that were not significantly different from the normal native cartilage in both the deep half and superficial half of interface repair tissue (P > .05). Conclusion: This study demonstrated that the anterolateral tibial osteotomy was a reasonable alternative for accessing centrolateral or posterolateral OLT for AOT with limited morbidity associated with the osteotomy. The evidence demonstrated adequate osteotomy and cartilaginous healing with improvement in functional outcome scores at mediumterm follow-up. Level of Evidence: Level IV, retrospective case series. [ABSTRACT FROM AUTHOR]

The article presents a retrospective case series of patients undergoing a flatfoot surgery with spring (calcaneonavicular) ligament reconstruction using a peroneus longus tendon autograft. In all thirteen cases, lateral column lengthening failed to correct talonvavicular deformity. Research found that spring ligament reconstruction resulted in few complications.

The authors discuss the findings of their research into the assessment of the cause of lateral hindfoot pain in flatfoot deformity. They compared assessments of patients with flatfoot deformity who either had no pain or experienced lateral hindfoot pain. They determined that weightbearing multiplanar imaging was a reliable method of assessing arthrosis and impingement in patients with flatfoot deformity and lateral pain.

Introduction/Purpose: The purpose of this study is to evaluate the effectiveness of Extracellular Matrix Cartilage Allograft (EMCA) as an adjuvant to bone marrow stimulation (BMS) compared to BMS alone in the treatment of osteochondral lesions of the talus (OLT). Methods: A retrospective cohort study comparing patients treated with BMS with EMCA (BMS-EMCA group) and BMS alone (BMS group) between 2013 and 2019 was undertaken. Clinical outcome was evaluated with the Foot and Ankle Outcome Score (FAOS) pre-and postoperatively. Postoperative MRIs were evaluated using a modified Magnetic Resonance Observation of Cartilage Tissue (MOCART) score. Comparisons between groups were made with the Man-Whitney U test for continuous variables and the ?2 or Fisher exact test for categorical variables. Results: Twenty-four patients underwent BMS with EMCA (BMS-EMCA group) and 24 patients underwent BMS alone (BMS group). The mean age was 40.8 years (range, 19 to 60 years) in BMS-EMCA group and 47.8 years (range, 24 to 60 years) in BMS group (p=0.060). The mean follow-up time was 20.0 months (range, 12-36 months) in BMS-EMCA group and 26.9 months (range, 12 to 55 months) in BMS group (p=0.031). Both groups showed significant improvements in all FAOS subscales. No significant differences between groups were found in all postoperative FAOS. The mean MOCART score in BMS-EMCA group was higher (76.3 vs 66.3), but not statistically significant (p=0.176). The MRI analysis showed that 87.5% of BMS-EMCA group had complete infill of the defect with repair tissue, however less than half (46.5%) of BMS group had complete infill (p=0.015). Conclusion: BMS with EMCA is an effective treatment strategy for the treatment of OLT and provides better cartilage infill in the defect on MRI. However, this did not translate to improved functional outcomes compared with BMS alone in the short-term. [ABSTRACT FROM AUTHOR]

Introduction/Purpose: Microfracture has been widely published as a treatment modality for osteochondral lesions of the talus (OLT). However, little is known about the outcome predictors following microfracture for smaller-sized OLT (<100mm2). This study sought to define the predictors of both clinical and magnetic resonance imaging (MRI) outcomes for small OLT treated with microfracture. Methods: A retrospective cohort study investigating patients who received arthroscopic microfracture for OLT (<10mm or 100mm2) between 2008 and 2017 were evaluated. Multivariate regression models were used to evaluate factors affecting postoperative Foot and Ankle Outcome Scores (FAOS) and Magnetic Resonance Observation of Cartilage Repair Tissue (MOCART) scores. Kaplan-Meier survival curves with log-rank test were constructed and endpoint was defined as the requirement of revision surgery. Results: Eighty-seven patients were included in the study. The mean follow-up time was 41.4+-28.7 months. The mean FAOS pain score significantly improved from 60.4+-14.8 preoperatively to 79.3+-12.8 at final follow-up (p<0.001). Patients with uncontainedtype OLT had an approximately 9-point worse FAOS pain score compared to contained-type OLT (p=0.036). Patients with cystic OLT also had an approximately 9-point worse pain FAOS compared to non-cystic OLT (p=0.026). Patients with larger lesion sizes had worse postoperative MOCART scores (p=0.012). Both Uncontained-Cyst and Uncontained-Noncyst groups had significantly worse FAOS pain than the Contained-Noncyst group (p<0.001, p=0.026). Survival rates in uncontained and contained lesions were 51.5% and 84.4%, respectively (p=0.616). Conclusion: Lesion uncontainment and the existence of cysts are independent predictors of poor clinical outcome following arthroscopic microfracture for smaller-sized OLT (<100mm2). [ABSTRACT FROM AUTHOR]

We describe a case of endovascular exclusion of a splenic artery aneurysm. The repair utilized a novel balloon-expandable stent-graft. [ABSTRACT FROM AUTHOR]

Purpose: To demonstrate the value of pretraining with millions of radiologic images compared with ImageNet photographic images on downstream medical applications when using transfer learning., Materials and Methods: This retrospective study included patients who underwent a radiologic study between 2005 and 2020 at an outpatient imaging facility. Key images and associated labels from the studies were retrospectively extracted from the original study interpretation. These images were used for RadImageNet model training with random weight initiation. The RadImageNet models were compared with ImageNet models using the area under the receiver operating characteristic curve (AUC) for eight classification tasks and using Dice scores for two segmentation problems., Results: The RadImageNet database consists of 1.35 million annotated medical images in 131 872 patients who underwent CT, MRI, and US for musculoskeletal, neurologic, oncologic, gastrointestinal, endocrine, abdominal, and pulmonary pathologic conditions. For transfer learning tasks on small datasets-thyroid nodules (US), breast masses (US), anterior cruciate ligament injuries (MRI), and meniscal tears (MRI)-the RadImageNet models demonstrated a significant advantage ( P < .001) to ImageNet models (9.4%, 4.0%, 4.8%, and 4.5% AUC improvements, respectively). For larger datasets-pneumonia (chest radiography), COVID-19 (CT), SARS-CoV-2 (CT), and intracranial hemorrhage (CT)-the RadImageNet models also illustrated improved AUC ( P < .001) by 1.9%, 6.1%, 1.7%, and 0.9%, respectively. Additionally, lesion localizations of the RadImageNet models were improved by 64.6% and 16.4% on thyroid and breast US datasets, respectively., Conclusion: RadImageNet pretrained models demonstrated better interpretability compared with ImageNet models, especially for smaller radiologic datasets. Keywords: CT, MR Imaging, US, Head/Neck, Thorax, Brain/Brain Stem, Evidence-based Medicine, Computer Applications-General (Informatics) Supplemental material is available for this article. Published under a CC BY 4.0 license.See also the commentary by Cadrin-Chênevert in this issue., Competing Interests: Disclosures of conflicts of interest: X.M. Member of Radiology: Artificial Intelligence trainee editorial board. Z.L. No relevant relationships. P.M.R. No relevant relationships. B.M. Consulting fees from Coridea. M.H. Portable knee MRI unit lent by Hyperfine to author’s institution, free of charge, for research purposes. A.D. No relevant relationships. A.J. No relevant relationships. C.C. No relevant relationships. K.E.L. Medical student research grant from Radiological Society of North America, payment made to NYU Langone; medical student scholarship from American College of Radiology, payment made to author. T.Y. No relevant relationships. Y.W. Grant support provided to author’s institution through NSF DMS-1752709 and NSF DMS-1720489 by the National Science Foundation. H.G. No relevant relationships. T.D. Radiologic data for the study was provided by RadImageNet and author is managing partner of RadImageNet. Z.A.F. Grants from Daiichi Sankyo, Amgen, Bristol Myers Squibb, and Siemens Healthineers; personal fees from Alexion, GlaxoSmithKline, and Trained Therapeutix Discovery; patents with Trained Therapeutix Discovery. Y.Y. Mount Sinai Seed Fund., (© 2022 by the Radiological Society of North America, Inc.)

Background: Ultrasound (US) is the most commonly used radiologic modality to identify and characterize thyroid nodules. Many nodules subsequently undergo fine needle aspiration to further characterize the nodule and determine appropriate treatment. The fine needle aspirate is most commonly classified using the Bethesda System for Reporting Thyroid Cytology (TBSRTC). It can sometimes be difficult to differentiate Bethesda class III lesions (atypia of undetermined significance/follicular lesion of undetermined significance) from Bethesda class IV, V and VI (malignant nodules). However, differentiation is important as clinical management differs between the two groups. The purpose of this study was to introduce machine learning methods to help radiologists differentiate Bethesda class III from Bethesda class VI, V and VI lesions., Methods: The authors collected 467 thyroid nodules with cytopathology results. US features were summarized using the 2017 ACR (American College of Radiology) Thyroid Imaging Reporting And Data System (TIRADS). Machine learning models [logistic regression, gradient boost, support vector machine (SVM), random forest and deep neural networks (DNN)] were created to classify Bethesda class III vs class IV/V/VI., Results: DNN outperformed other machine learning classifiers and obtained the highest accuracy and specificity to classify thyroid nodules as either Bethesda III or IV/V/VI nodules using multiple US features., Conclusions: Machine learning/deep learning approaches could help differentiate Bethesda III nodules from IV/V/VI using US features which may benefit treatment decisions., Competing Interests: Conflicts of Interest: The authors have no conflicts of interest to declare.

Competing Interests: Conflicts of Interest: The authors have no conflicts of interest to declare.

Background: The purpose of this study was to develop a predictive model that could accurately predict the malignancy of the pulmonary ground-glass nodules (GGNs) and the invasiveness of the malignant GGNs., Methods: The authors built two binary classification models that could predict the malignancy of the pulmonary GGNs and the invasiveness of the malignant GGNs., Results: Results of our developed model showed random forest could achieve 95.1% accuracy to predict the malignancy of GGNs and 83.0% accuracy to predict the invasiveness of the malignant GGNs., Conclusions: The malignancy and invasiveness of pulmonary GGNs could be predicted by random forest., Competing Interests: Conflicts of Interest: The authors have no conflicts of interest to declare.