AJR articles provide information and analysis on technical innovations and frontline research across several subspecialties, examining challenges relevant to today’s radiologists. Find these complete articles and more at AJRonline.org.
- PET/CT Plays Role in Lung Adenocarcinoma Management
Fluorodeoxyglucose PET (FDG PET) can be used to predict the histopathologic subtypes and growth patterns of early lung adenocarcinoma. “FDG PET, combined with high-resolution CT (HRCT), has value for predicting invasive histopathologic subtypes, but there was no significance for predicting invasive growth patterns,” clarified authors Xiaoliang Shao and Xiaonan Shao from the department of nuclear medicine at Soochow University in Changzhou, China. The team’s retrospective analysis was conducted on the PET/CT data on ground-glass nodules (GGNs) resected from patients with stage IA lung adenocarcinoma, evaluating the efficacy of PET maximum standardized uptake value (SUVmax) combined with HRCT signs in prediction of histopathologic subtype and growth pattern of lung adenocarcinoma. Although SUVmax measured significantly higher in GGNs with invasive HRCT signs, the diameter of GGN, as well as the attenuation value differential between ground-glass components and adjacent lung tissues, were independent predictors of FDG uptake by GGNs. Additionally, SUVmax was higher in invasive adenocarcinoma than in adenocarcinoma in situ (AIS)–minimally invasive adenocarcinoma (MIA), with SUVmax 2.0 the optimal cutoff value for differentiation. Acinar-papillary adenocarcinoma had a higher SUVmax than lepidic adenocarcinoma, with SUVmax 1.4 the optimal cutoff value for differentiation. “In stage IA lung adenocarcinoma characterized by GGNs, the SUVmax of GGNs with invasive CT features was high,” Xiaoliang Shao wrote, adding that HRCT can be used in diagnosing the subtypes of lung adenocarcinoma. “However, it cannot be used to differentiate different growth patterns of lung adenocarcinomas.” As Xiaonan Shao concluded: “The efficacy of FDG PET SUVmax in differentiating lung adenocarcinoma subtypes is similar to that of HRCT signs, however, the diagnostic efficiency of FDG PET combined with HRCT is significantly higher than that of each imaging technique alone.”
- MRI Predicts Shoulder Stiffness for Rotator Cuff Tears
Two MRI findings—joint capsule edema and thickness at the axillary recess, specifically—proved useful in predicting stiff shoulder in patients with rotator cuff tears. Studying 106 patients with small to large (≤ 5 cm) full-thickness rotator cuff tears, in addition to joint capsule edema and thickness in the axillary recess, Yoon Yi Kim of Korea’s Veterans Health Service Medical Center assessed obliteration of the subcoracoid fat triangle, fatty degeneration of the torn rotator cuff muscle, and degree of retraction. Tear size and location were determined by MRI findings and operative report, while associations between MRI findings and preoperative passive range of motion (ROM) were evaluated with simple and multiple linear regression analyses and proportional odds logistic regression analysis. As Kim and colleagues wrote: “There was a significant, negative linear correlation between limited ROM at forward elevation and thickness of the joint capsule in the glenoid portion of the axillary recess (p = 0.018), external rotation and joint capsule edema in the humeral portion of the axillary recess (p = 0.011), and internal rotation and joint capsule edema in the glenoid portion of the axillary recess (p = 0.007).” Fatty degeneration (p = 0.003) was an independent predictor of limited ROM on internal rotation. Meanwhile, male sex (p = 0.041) and posterosuperior rotator cuff tear (p = 0.030) were independent predictors of shoulder ROM on external rotation. “This study is important,” Kim et al. noted, “because it is the first to highlight joint capsule abnormality on MRI as a factor associated with stiff shoulder in patients with full-thickness rotator cuff tears.”
- New CT Scoring Criteria for Timely Diagnosis and Treatment of COVID-19
Updated CT scoring criteria that considers lobe involvement, as well as changes in CT findings, could quantitatively and accurately evaluate the progression of coronavirus disease (COVID-19) pneumonia. “The earlier that COVID-19 is diagnosed and treated, the shorter the time to disease resolution and the lower the highest and last CT scores are,” concluded lead author Guoquan Huang of Wuhu Second People’s Hospital in China. Assigning CT scores to 25 patients according to CT findings and lung involvement, Huang and colleagues recorded the time from symptom onset to diagnosis and treatment for each patient. Patients with COVID-19 were divided into two groups: (patients for whom this interval was ≤ 3 days) and group 2 (those for whom the interval was > 3 days). Using a Lorentzian line-shape curve to show the variation tendency during treatment, the fitted tendency curves for group 1 and group 2 were significantly different. Peak points showed that the estimated highest CT score was 10 and 16 for each group, respectively, and the time to disease resolution was 6 and 13 days, respectively. The Mann-Whitney test showed that the last CT scores were lower for group 1 than for group 2 (p = 0.025), although the chi-square test found no difference in age and sex between the groups. The time from symptom onset to diagnosis and treatment had a positive correlation with the time to disease resolution (r = 0.93; p = 0.000), as well as with the highest CT score (r = 0.83; p = 0.006). “Sequential chest CT examinations enable qualitative investigation of alterations in COVID-19 infection during the course of treatment,” Huang explained. Because previously proposed CT scoring criteria regarding lobe involvement gave no consideration to changes in CT features (i.e., the change from observation of GGO to a crazy-paving pattern and then consolidation), Huang et al. suggest that such a rubric is not sufficiently accurate to assess the progression of pneumonia. “In the present study,” wrote Huang, “we propose a new version of CT scoring criteria that considers both lobe involvement and changes in CT findings, in an attempt to more comprehensively evaluate COVID-19 pneumonia on sequential chest CT examinations.”
- Pediatric Coronavirus Disease (COVID-19) Pneumonia Radiography, CT Findings Included in Review of Five New Lung Disorders
Although the clinical symptoms of new pediatric lung disorders such as severe acute respiratory syndrome (SARS), swine-origin influenza A (H1N1), Middle East respiratory syndrome (MERS), e-cigarette or vaping product use–associated lung injury (EVALI), and coronavirus disease (COVID-19) pneumonia may be nonspecific, some characteristic imaging findings have emerged or are currently emerging. “Although there are some overlapping imaging features of these disorders,” wrote first author Alexandra M. Foust of Boston Children’s Hospital and Harvard Medical School, “careful evaluation of the distribution, lung zone preference, and symmetry of the abnormalities with an eye for a few unique differentiating imaging features, such as the halo sign seen in COVID-19 and subpleural sparing and the atoll sign seen in EVALI, can allow the radiologist to offer a narrower differential diagnosis in pediatric patients, leading to optimal patient care.” At most institutions, whereas the first imaging study performed in patients with clinically suspected COVID-19 is chest radiography, Foust and colleagues’ review of the clinical literature found that studies on chest radiography findings in patients with COVID-19 were relatively scarce. Regarding the limited studies of pediatric patients with COVID-19, Foust et al. noted chest radiography “may show normal findings; patchy bilateral ground-glass opacity (GGO), consolidation, or both; peripheral and lower lung zone predominance.” Similarly, while the literature describing chest CT findings in patients with COVID-19 are more robust than those describing chest radiography findings, only a few articles have reported CT findings of COVID-19 in children. A study of 20 pediatric patients with COVID-19 reported that the most frequently observed abnormalities on CT were subpleural lesions (100% of patients), unilateral (30%) or bilateral (50%) pulmonary lesions, GGO (60%), and consolidation with a rim of GGO surrounding it, also known as the halo sign (50%). The authors of this AJR article also pointed to a smaller study of five pediatric patients with COVID-19, where investigators reported modest patchy GGO, one with peripheral subpleural involvement, in three patients that resolved on follow-up CT examination.
- Review of COVID-19 Studies Cautions Against Chest CT for Coronavirus Diagnosis
To date, the radiology literature on coronavirus disease (COVID-19) pneumonia has consisted of limited retrospective studies that do not substantiate the use of CT as a diagnostic test for COVID-19. “This is not to say these studies are not valuable,” maintained lead investigator Constantine A. Raptis of Washington University in Saint Louis. As Raptis, Travis S. Henry of the University of California-San Francisco, and nine co-authors from six institutions across the United States noted of the most frequently cited studies on the subject, reporting the various CT features of COVID-19 pneumonia remains “an important first step” in helping radiologists identify patients who may have COVID-19 in the appropriate clinical environment. “However,” they continue, “test performance and management issues arise when inappropriate and potentially overreaching conclusions regarding the diagnostic performance of CT for COVID-19 pneumonia are based on low-quality studies with biased cohorts, confounding variables, and faulty design characteristics.” Because misdiagnosing even a single patient (i.e., obtaining a false-negative finding) could result in large outbreaks among future contacts, understanding the potential effects of selection bias is important in determining sensitivity. As Raptis and colleagues explained, “if a study cohort contains patients who are more likely to have a true-positive finding and less likely to have a false-negative finding, sensitivity will be overestimated.” The specificity and positive predictive value of a laboratory test—in the case of COVID-19, reverse transcription–polymerase chain reaction (RT-PCR)—are based on its ability to limit false-positive findings. Acknowledging false-positive RT-PCR results are possible, Raptis, Henry, et al. maintained they are often caused by contamination and are likely insignificant in the setting of assays for COVID-19. CT, on the other hand, does not test for singular features unique to the disease, and even those features most characteristic of COVID-19 pneumonia—peripheral, bilateral ground-glass opacities typically in the lower lobes—have been reported in a number of other conditions, both infectious and noninfectious. Finally, Raptis and colleagues addressed the hazards of wide deployment of CT: overuse of hospital resources, including the use of PPE already limited in availability but required to safely perform CT studies; clustering of affected and nonaffected patients in the radiology department, increasing the risk of disease transmission among imaging staff. “At present,” the authors of this AJR article concluded, “CT should be reserved for evaluation of complications of COVID-19 pneumonia or for assessment if alternative diagnoses are suspected.”
- Low Back Pain Accounts for a Third of New Emergency Imaging in the U.S.
The use of imaging for the initial evaluation of patients with low back pain in the emergency department (ED) continues to occur at a high rate—one in three new emergency visits for low back pain in the United States. “Although there has been a modest decline,” wrote Jina Pakpoor of the University of Pennsylvania, “in 2016, approximately one in three patients still continued to receive imaging in the ED. Further, significant geographic variation exists between differing states and regions of the United States.” Pakpoor and colleagues identified ED visits for patients with low back pain billed to insurance by querying IBM’s Commercial Claims and Encounters Marketscan research database for patients 18–64 years old. Excluding patients with concomitant encounter diagnoses suggesting trauma, as well as those with previous visits for back pain, Current Procedural Terminology codes were used to identify three imaging modalities: radiography, CT, and MRI. Of the 134,624 total encounters meeting Pakpoor’s inclusion criteria, imaging was obtained in 44,405 (33.7%) visits and decreased from 34.4% to 31.9% between 2011 and 2016 (odds ratio per year, 0.98 [95% CI, 0.98– 0.99]; p < 0.001). During the five-year study period, 30.9% of patients underwent radiography, 2.7% of patients underwent CT, and 0.8% of patients underwent MRI for evaluation of low back pain. Imaging utilization varied significantly by geographic region (p < 0.001), with patients in the southern U.S. undergoing 10% more imaging than patients in the western U.S. Acknowledging further research is necessary “to understand the underlying reasons for persistent use of potentially unwarranted imaging in the emergency setting,” as Pakpoor concluded, “our results indicate that the use of imaging for the evaluation of patients with low back pain in the ED is moderately declining but continues to occur at an overall high rate.”
- Deep Learning Differentiates Small Renal Masses on Multiphase CT
A deep learning method with a convolutional neural network (CNN) can support the evaluation of small solid renal masses in dynamic CT images with acceptable diagnostic performance. Between 2012 and 2016, researchers at Japan’s Okayama University studied 1,807 image sets from 168 pathologically diagnosed small (≤ 4 cm) solid renal masses with four CT phases—unenhanced, corticomedullary, nephrogenic, and excretory—in 159 patients. Masses were classified as malignant (n = 136) or benign (n = 32) using a 5-point scale, and this dataset was then randomly divided into five subsets. As first AJR author, Takashi Tanaka explained, “four were used for augmentation and supervised training (48,832 images), and one was used for testing (281 images).” Utilizing the Inception-v3 architecture CNN model, the AUC for malignancy and accuracy at optimal cutoff values of output data were evaluated in six different CNN models. Finding no significant size difference between malignant and benign lesions, Tanaka’s team did find that the AUC value of the corticomedullary phase was higher than that of other phases (corticomedullary vs excretory, p = 0.022). Additionally, the highest accuracy (88%) was achieved in the corticomedullary phase images. Multivariate analysis revealed that the CNN model of corticomedullary phase was a significant predictor for malignancy, “compared with other CNN models, age, sex, and lesion size.”
- A Step Toward Automated Triage of Thyroid Cancer
A Stanford University team has developed a quantitative framework able to sonographically differentiate between benign and malignant thyroid nodules at a level comparable to that of expert radiologists, which may prove useful for establishing a fully automated system of thyroid nodule triage. Alfiia Galimzianova et al. retrospectively collected ultrasound images of 92 biopsy-confirmed nodules, which were annotated by two expert radiologists using the American College of Radiology’s Thyroid Imaging Reporting and Data System (TI-RADS). In the researchers’ framework, nodule features of echogenicity, texture, edge sharpness, and margin curvature properties were analyzed in a regularized logistic regression model to predict nodule malignancy. Authenticating their method with leave-one-out cross-validation, the Stanford team used ROC AUC, sensitivity, and specificity to compare the framework’s results with those obtained by six expert annotation-based classifiers. Galimzianova et al. noted that the AUC of the proposed framework measured 0.828 (95% CI, 0.715–0.942)—“greater than or comparable to that of the expert classifiers”—whose AUC values ranged from 0.299 to 0.829 (p = 0.99). Additionally, in a curative strategy at sensitivity of 1, use of the framework could have avoided biopsy in 20 of 46 benign nodules—statistically significantly higher than three expert classifiers. In a conservative strategy at specificity of 1, the framework could have helped to identify 10 of 46 malignancies— statistically significantly higher than five expert classifiers. “Our results confirm the ultimate feasibility of computer-aided diagnostic systems for thyroid cancer risk estimation,” concluded Galimzianova. “Such systems could provide second-opinion malignancy risk estimation to clinicians and ultimately help decrease the number of unnecessary biopsies and surgical procedures.”
- What to Know About Frostbite and Amputation
An AJR article reviewing various techniques and clinical management paradigms to treat severe frostbite injuries— especially relevant for interventional radiologists showed promising results using both intraarterial (IA) and IV tPA (tissue plasminogen activator) to reduce amputation. “Severe frostbite injuries can lead to devastating outcomes with loss of limbs and digits, yet clinical management continues to consist primarily of tissue rewarming, prolonged watchful waiting, and often delayed amputation,” wrote Boston Medical Center radiologists John Lee and Mikhail Higgins. A search of the literature by Lee and Higgins yielded 157 publications. After manually screening for inclusion criteria of case reports, case series, cohort studies, and randomized prospective studies that reported the use of tPA to treat severe frostbite injuries, 16 qualified for review. Lee and Higgins’ analyzed series included 209 patients with 1,109 digits at risk of amputation treated with IA or IV tPA—116 and 77 patients, respectively. A total 926 at-risk digits were treated with IA tPA and resulted in amputation of 222 digits, for a salvage rate of 76%. Twenty-four of 63 patients underwent amputation after IV tPA, resulting in a 62% salvage rate. Both digital subtraction angiography and triple-phase bone scan were utilized for initial imaging evaluation of patients with severe frostbite injuries. Additional concurrent treatment included therapeutic heparin at 500 U/h, warfarin with target international normalized ratio of 2:3, nonsteroidal antiinflammatory drugs, pain management, and light dressings with topical antimicrobial agents. “For many years,” Lee and Higgins concluded, “the axiom ‘frostbite in January, amputate in July’ was an accurate description of the common outcome in frostbite injuries. Through a meta-analysis of thrombolytic therapy in the management of severe frostbite, this article provides a useful guideline for interventional radiologists, including a suggested protocol, inclusion and exclusion criteria, and potential complications.”
- Smartphone, Laptop Prove Reliable and Accurate for Acute Stroke Decision
Mobile devices proved both reliable and accurate for the clinical decision to administer IV thrombolysis in patients with acute stroke, according to an ahead-of-print article in the April issue of AJR. To assess reliability and accuracy of IV thrombolysis recommendations made after interpretation of head CT images of patients with acute stroke symptoms displayed on smartphone or laptop reading systems—compared with those made after interpretation of images displayed on a medical workstation monitor—Antonio J. Salazar at the University of Los Andes in Bogotá, Columbia utilized a factorial design with 188 patients, four radiologists, and three reading systems to produce 2,256 interpretations. To evaluate reliability, Salazar and colleagues calculated the intraobserver and interobserver agreements using the intraclass correlation coefficient (ICC) and five interpretation variables: hemorrhagic lesions, intraaxial neoplasm, stroke dating (acute, subacute, chronic), hyperdense arteries, and infarct size assessment. Accuracy equivalence tests were also performed for the IV thrombolysis recommendation; for this variable, sensitivity, specificity, and ROC curves were evaluated. Good or very good intraobserver agreements were observed for all the variables. Specifically, for those variables required to establish contraindications for IV thrombolysis, the agreements were ranked as very good. “This finding is important,” wrote Salazar et al., “because it reflects the good performance of mobile devices to evaluate the most significant imaging variables for clinical decisions.” For IV thrombolysis recommendation, the main subject of this evaluation, the interobserver agreements for the three reading systems were ranked as very good (ICC > 0.88). Similarly, very good intraobserver agreements were observed for all comparisons (ICC > 0.84). The AUC values (0.83–0.84) and sensitivities (0.94–0.95) for IV thrombolysis recommendation were equivalent among all the reading systems at a 5% equivalent threshold. As a unique assessment of imaging-based recommendations for the administration of IV recombinant tissue plasminogen activator based on unenhanced brain CT scans, Salazar also noted: “These results constitute a strong foundation for the development of mobilebased telestroke services because they increase neuroradiologist availability and the possibility of using reperfusion therapies in resource-limited countries.”