Assessing the Potential of a Deep Learning Tool to Improve Fracture Detection by Radiologists and Emergency Physicians on Extremity Radiographs.

RATIONALE AND OBJECTIVES: To evaluate the standalone performance of a deep learning (DL) based fracture detection tool on extremity radiographs and assess the performance of radiologists and emergency physicians in identifying fractures of the extremities with and without the DL aid. MATERIALS AND METHODS: The DL tool was previously developed using 132,000 appendicular skeletal radiographs divided into 87% training, 11% validation, and 2% test sets. Stand-alone performance was evaluated on 2626 de-identified radiographs from a single institution in Ohio, including at least 140 exams per body region. Consensus from three US board-certified musculoskeletal (MSK) radiologists served as ground truth. A multi-reader retrospective study was performed in which 24 readers (eight each of emergency physicians, non-MSK radiologists, and MSK radiologists) identified fractures in 186 cases during two independent sessions with and without DL aid, separated by a one-month washout period. The accuracy (area under the receiver operating curve), sensitivity, specificity, and reading time were compared with and without model aid. RESULTS: The model achieved a stand-alone accuracy of 0.986, sensitivity of 0.987, and specificity of 0.885, and high accuracy (> 0.95) across stratification for body part, age, gender, radiographic views, and scanner type. With DL aid, reader accuracy increased by 0.047 (95% CI: 0.034, 0.061; p = 0.004) and sensitivity significantly improved from 0.865 (95% CI: 0.848, 0.881) to 0.955 (95% CI: 0.944, 0.964). Average reading time was shortened by 7.1 s (27%) per exam. When stratified by physician type, this improvement was greater for emergency physicians and non-MSK radiologists. CONCLUSION: The DL tool demonstrated high stand-alone accuracy, aided physician diagnostic accuracy, and decreased interpretation time.

Play in the reading room: Utilizing soft modeling compound to teach musculoskeletal anatomy and pathology.

PROBLEM DESCRIPTION: Musculoskeletal (MSK) anatomy and pathology from a radiology perspective can be difficult to conceptualize and understand due to the challenge of visualizing 3D structures in stacks of 2D imaging. Consequently, trainees may benefit from inexpensive methods that can help trainees better visualize MSK anatomy and pathology. The purpose of this study is to provide proof of concept for inexpensive methodology to help learners such as radiology residents quickly and inexpensively understand musculoskeletal anatomy and pathology. This can help trainees become better at applying musculoskeletal knowledge to clinical practice. INSTITUTIONAL METHODOLOGY: Soft-modeling compounds such as Play-Doh® was utilized in a variety of colors with pottery tools to recreate 3D models of challenging MSK anatomy and pathology for trainees. Qualitative feedback from the residents was collected. RESULTS: Eighteen different pathological conditions across six major bone structures were modeled with a soft modeling compound. Residents qualitatively identified the experience as educational in terms of helping them better understand MSK pathology and positive in terms of making learning fun, less stressful, and memorable due to uniqueness of the learning modality. Residents report challenges modeling complex anatomical features and pathology via this methodology. CONCLUSION: Radiology residents and other learners can enhance their knowledge of musculoskeletal anatomy and pathology via utilization of inexpensive soft modeling compounds. This may offer a cheaper and more time sensitive alternative to current 3-dimensional hardware and software technologies being developed for educational purposes. Additional work needs to be done to examine the utility of this methodology across larger and diverse groups of learners.

Sleeve gastrectomy morphology and long-term weight-loss and gastroesophageal reflux disease outcomes.

BACKGROUND: The relationship between sleeve gastrectomy (SG) morphology and long-term weight-loss and gastroesophageal reflux disease (GERD) outcomes is unknown. METHODS: All patients (n = 268) undergoing SG performed by 3 surgeons at a single academic institution from January 1, 2010 to December 31, 2012 were included. Long-term weight-loss and GERD outcomes were available for 90 patients which were incorporated in analyses. SG morphology was determined from postoperative day 1 upper gastrointestinal series (UGIS) available from 50 patients. Images were independently categorized using previously published methodology as Dumbbell (38%), Lower Pouch (22%), Tubular (26%), or Upper Pouch (14%) by Radiologist and Surgeon. Radiologist categorization was used when disagreement occurred (8%). Univariable analyses were conducted to explore potential associations between SG morphology, weight loss, and GERD outcomes. RESULTS: Follow-up was 8.2 ± 0.9 years. Population characteristics included age of 45.1 ± 10.8 years, female sex in 83.3%, and hiatal hernia repair (HHR) performed at index SG in 17.8%. Surgeons did not preferentially achieve a specific SG morphology. Changes from preoperative obesity and associated diseases comprised body mass index (BMI) (49.5 ± 7.6 vs. 39.2 ± 9.4 kg/m2; p < 0.0001), diabetes mellitus (30.0 vs. 12.2%; p = 0.0006), hypertension (70.0 vs. 54.4%; p = 0.0028), hyperlipidemia (42.2 vs. 24.2%;p = 0.0017), obstructive sleep apnea (41.1 vs. 15.6%; p < 0.0001), osteoarthritis (48.9 vs. 13.3%; p < 0.0001), back pain (46.5 vs. 28.9%; p = 0.0035), and medications (4.8 ± 3.3 vs. 3.7 ± 3.5; p < 0.0001). Dumbbell SG morphology was associated with lesser reduction in BMI at follow-up (--6.8 ± 7.2 vs. -12.4 ± 8.3 kg/m2; p = 0.0196) while greater BMI change was appreciated with Lower Pouch SG shape (-16.9 ± 9.9 vs. -8.4 ± 6.8 kg/m2; p = 0.0017). GERD was more prevalent at follow-up than baseline (67.8 vs. 47.8%; p < 0.0001). GERD-specific outcomes included de novo (51.1%), persistent (27.9%), worsened (58.1%), and resolved (14.0%) disease. Ten patients underwent reoperation for refractory GERD with SG morphology corresponding to Dumbbell (n = 5) and Upper Pouch (n = 1) for those with available UGIS. Univariable analyses showed that patients with GERD experienced a larger reduction in BMI compared with patients without GERD (-11.8 ± 7.7 vs. -7.0 ± 5.1 kg/m2; p = 0.0007). Patient age, surgeon, morphology category, and whether a HHR was done at index SG were not associated with the presence of any, de novo, or worsened GERD. Female sex was associated with worsened GERD (96.0 vs. 4.0%; p = 0.0455). Type of calibration device, distance from staple line to pylorus, and whether staple line reinforcement was used were not associated with SG morphology classification. CONCLUSION: This is the first study assessing the impact of SG morphology on long-term weight loss and GERD. Our data suggest an association between SG morphology and long-term weight loss but not with GERD outcomes. Current technical standards may be limited in reproducing the same SG morphology. This information may help guide the technical optimization and standardization of SG. Surgeons did not favor a specific SG morphology (1). Our results signal to a relationship between radiographic assessment of SG morphology and long-term weight-loss outcomes with Dumbbell classification correlated with lesser reduction in BMI (2a) and Lower Pouch morphology associated with superior weight loss (2b). SG, sleeve gastrectomy; BMI, body mass index.

Ischial osteochondroma as an unusual source of pregnancy-related sciatic pain: a case report.

BACKGROUND: While most cases of sciatica result from degenerative conditions of the low back, some cases result from conditions of the hip and pelvic region. Sciatica developing in relation to pregnancy or labor also presents unique considerations. CASE PRESENTATION: A 37-year-old African American woman with a history of hypertension and polycystic ovary syndrome presented to a chiropractor at a hospital-based outpatient clinic with a seven-week history of low back pain with radiation into the right lower extremity which began during labor. The chiropractor performed a brief trial of care, yet when the patient's symptoms worsened, ordered lumbar spine radiographs, followed by lumbar magnetic resonance imaging (MRI), which were both normal. The chiropractor then ordered hip radiographs, which were suggestive of ischial osteochondroma, and referred the patient to an orthopedic oncologist. MRI findings were compatible with an osteochondroma with associated adventitial bursitis and mass effect on the sciatic nerve. The patient initially chose conservative management with bursa aspiration and therapeutic injection. Despite initial relief, there was eventual return of symptoms. The patient elected to undergo surgical removal, with a positive outcome. CONCLUSION: The key distinguishing features that led to a diagnosis of osteochondroma in this case included attention to the patient-reported symptoms and history, worsening of symptoms despite conservative care, and lack of explanatory findings on lumbar imaging. This case highlights the benefit of evaluating the hip and pelvis when the clinical features of sciatica cannot be ascribed to a lumbar etiology. This case also illustrates the role of a chiropractor working in an integrative health system to facilitate timely imaging and referrals to resolve a challenging diagnosis.

Moderate sedation: what radiologists need to know.

OBJECTIVE: The number of procedures conducted within the radiology department using moderate sedation is rising. Anecdotal evidence suggests that there is an inconsistency in the education of radiology trainees regarding moderate sedation. CONCLUSION: The purpose of this article is to give an overview and concise summary of the use of moderate sedation as put forth by the American Society of Anesthesia, American College of Radiology, and Society of Interventional Radiology. In addition, we will review the commonly used medications for moderate sedation and their reversal agents.

Diagnostic performance of dual-time 18F-FDG PET in the diagnosis of pulmonary nodules: a meta-analysis.

RATIONALE AND OBJECTIVE: Perform a comprehensive meta-analysis evaluating the diagnostic performance of dual time point deoxy-2-[(18)F]fluoro-D-glucose positron emission tomography (FDG-PET) in the diagnosis of pulmonary nodules. MATERIALS AND METHODS: MEDLINE, EMBASE, and PUBMED were queried between January 2000 and January 2011. Studies were included if they: 1) used dual time point FDG-PET as a diagnostic test for pulmonary nodules, 2) used pathology or clinical follow-up as the reference standard, and 3) reported absolute number of true-positive (TP), true-negative (TN), false-positive (FP), and false-negative (FN) results or stated sufficient data to derive these values. Summary sensitivity (SN), summary specificity (SP), positive and negative likelihood ratios (LR+) and (LR-), and diagnostic odds ratio were calculated. Heterogeneity of the results was assessed using Forest plots and the value of inconsistency index (I(2)). RESULTS: Inclusion criteria were fulfilled by 10 articles with a total of 816 patients and 890 pulmonary nodules. The summary sensitivity was 85% (82%-89% at 95% confidence interval [CI]) and summary specificity was 77% (CI: 72%-81%), with a LR+ of 2.7 (CI: 1.4-5.2) and a LR- of 0.26 (CI: 0.14-0.49). Diagnostic odds ratio was 11 (CI: 3.8-32.2). Significant heterogeneity was found in the sensitivity (I(2) = 77%) and specificity (90.3%). CONCLUSION: Dual time point FDG-PET demonstrates similar sensitivity and specificity to single time point FDG-PET in the diagnosis of pulmonary nodules. The additive value of the dual time point FDG-PET is questionable, primarily because of the significant overlap of benign and malignant nodule FDG-PET characteristics and lack of consensus criteria for quantitative thresholds to define nodules as malignant.

Diagnostic performance of magnetic resonance imaging in the detection of appendicitis in adults: a meta-analysis.

RATIONALE AND OBJECTIVES: Perform a meta-analysis evaluating the diagnostic performance of magnetic resonance imaging (MRI) for the diagnosis of acute appendicitis. MATERIALS AND METHODS: MEDLINE and EMBASE were queried between January 1995 and December 2009. Prospective and retrospective studies were included if they: used MRI as a diagnostic test for appendicitis, used pathology or clinical follow-up as the reference standard, and reported absolute number of true-positive, true-negative, false-positive, and false-negative results, or stated sufficient data to derive these values. Summary sensitivity, summary specificity, positive and negative likelihood ratios (LR+) and (LR-), and diagnostic odds ratio were calculated. Heterogeneity of the results was assessed using Forest plots and the value of inconsistency index (I(2)). RESULTS: The inclusion criteria were fulfilled by eight articles with a total of 363 patients (mean age 26.9 ± 7.2 years; 86.2 % female). The appendix was not found in eight patients, with one article not reporting such data. The summary sensitivity was 97% (92%-99% at 95% confidence interval [CI]) and summary specificity was 95% (CI: 94%-99%), with a LR+ of 16.3 (CI: 9.1-29.1) and a LR- of 0.09 (CI: 0.04-0.197). Diagnostic odds ratio was 299.85 (CI: 97.5-921.61). No heterogeneity was found in the sensitivity (I(2) = 0.0, P = .4589). Minimal heterogeneity was found in the specificity (I(2) = 21.9%, P = .2553). CONCLUSION: MRI appears promising in the evaluation of acute appendicitis, although larger future studies are warranted to confirm the results.