Can AI distinguish a bone radiograph from photos of flowers or cars? Evaluation of bone age deep learning model on inappropriate data inputs.

OBJECTIVE: To evaluate the behavior of a publicly available deep convolutional neural network (DCNN) bone age algorithm when presented with inappropriate data inputs in both radiological and non-radiological domains. METHODS: We evaluated a publicly available DCNN-based bone age application. The DCNN was trained on 12,612 pediatric hand radiographs and won the 2017 RSNA Pediatric Bone Age Challenge (concordance of 0.991 with radiologist ground-truth). We used the application to analyze 50 left-hand radiographs (appropriate data inputs) and seven classes of inappropriate data inputs in radiological (i.e., chest radiographs) and non-radiological (i.e., image of street numbers) domains. For each image, we noted if (1) the application distinguished between appropriate and inappropriate data inputs and (2) inference time per image. Mean inference times were compared using ANOVA. RESULTS: The 16Bit Bone Age application calculated bone age for all pediatric hand radiographs with mean inference time of 1.1 s. The application did not distinguish between pediatric hand radiographs and inappropriate image types, including radiological and non-radiological domains. The application inappropriately calculated bone age for all inappropriate image types, with mean inference time of 1.1 s for all categories (p = 1). CONCLUSION: A publicly available DCNN-based bone age application failed to distinguish between appropriate and inappropriate data inputs and calculated bone age for inappropriate images. The awareness of inappropriate outputs based on inappropriate DCNN input is important if tasks such as bone age determination are automated, emphasizing the need for appropriate oversight at the data input and verification stage to avoid unrecognized erroneous results.

Penis Transplantation: First US Experience.

OBJECTIVE: We describe the first successful penis transplant in the United States in a patient with a history of subtotal penectomy for penile cancer. BACKGROUND: Penis transplantation represents a new paradigm in restoring anatomic appearance, urine conduit, and sexual function after genitourinary tissue loss. To date, only 2 penis transplants have been performed worldwide. METHODS: After institutional review board approval, extensive medical, surgical, and radiological evaluations of the patient were performed. His candidacy was reviewed by a multidisciplinary team of surgeons, physicians, psychiatrists, social workers, and nurse coordinators. After appropriate donor identification and recipient induction with antithymocyte globulin, allograft procurement and recipient preparation took place concurrently. Anastomoses of the urethra, corpora, cavernosal and dorsal arteries, dorsal vein, and dorsal nerves were performed, and also inclusion of a donor skin pedicle as the composite allograft. Maintenance immunosuppression consisted of mycophenolate mofetil, tacrolimus, and methylprednisolone. RESULTS: Intraoperative, the allograft had excellent capillary refill and strong Doppler signals after revascularization. Operative reinterventions on postoperative days (PODs) 2 and 13 were required for hematoma evacuation and skin eschar debridement. At 3 weeks, no anastomotic leaks were detected on urethrogram, and the catheter was removed. Steroid resistant-rejection developed on POD 28 (Banff I), progressed by POD 32 (Banff III), and required a repeat course of methylprednisolone and antithymocyte globulin. At 7 months, the patient has recovered partial sensation of the penile shaft and has spontaneous penile tumescence. Our patient reports increased overall health satisfaction, dramatic improvement of self-image, and optimism for the future. CONCLUSIONS: We have shown that it is feasible to perform penile transplantation with excellent results. Furthermore, this experience demonstrates that penile transplantation can be successfully performed with conventional immunosuppression. We propose that our successful penile transplantation pilot experience represents a proof of concept for an evolution in reconstructive transplantation.

Current Applications and Future Impact of Machine Learning in Radiology.

Recent advances and future perspectives of machine learning techniques offer promising applications in medical imaging. Machine learning has the potential to improve different steps of the radiology workflow including order scheduling and triage, clinical decision support systems, detection and interpretation of findings, postprocessing and dose estimation, examination quality control, and radiology reporting. In this article, the authors review examples of current applications of machine learning and artificial intelligence techniques in diagnostic radiology. In addition, the future impact and natural extension of these techniques in radiology practice are discussed.

Implementation of Digital Awareness Strategies to Engage Patients and Providers in a Lung Cancer Screening Program: Retrospective Study.

BACKGROUND: Lung cancer is the leading cause of cancer-related deaths in the United States. Despite mandated insurance coverage for eligible patients, lung cancer screening rates remain low. Digital platforms, including social media, provide a potentially valuable tool to enhance health promotion and patient engagement related to lung cancer screening (LCS). OBJECTIVE: The aim was to assess the effectiveness of LCS digital awareness campaigns on utilization of low-dose computed tomography (LDCT) and visits to institutional online educational content. METHODS: A pay-per-click campaign utilizing Google and Facebook targeted adults aged 55 years and older and caregivers aged 18 years and older (eg, spouses, adult children) with LCS content during a 20-week intervention period from May to September 2016. A concurrent pay-per-click campaign using LinkedIn and Twitter targeted health care providers with LCS content. Geographic target radius was within 60 miles of an academic medical center. Social media data included aggregate demographics and click-through rates (CTRs). Primary outcome measures were visits to institutional Web pages and scheduled LDCT exams. Study period was 20 weeks before, during, and after the digital awareness campaigns. RESULTS: Weekly visits to the institutional LCS Web pages were significantly higher during the digital awareness campaigns compared to the 20-week period prior to implementation (mean 823.9, SD 905.8 vs mean 51, SD 22.3, P=.001). The patient digital awareness campaign surpassed industry standard CTRs on Google (5.85%, 1108/18,955 vs 1.8%) and Facebook (2.59%, 47,750/1,846,070 vs 0.8%). The provider digital awareness campaign surpassed industry standard CTR on LinkedIn (1.1%, 630/57,079 vs 0.3%) but not Twitter (0.19%, 1139/587,133 vs 0.25%). Mean scheduled LDCT exam volumes per week before, during, and after the digital awareness campaigns were 17.4 (SD 7.5), 20.4 (SD 5.4), and 26.2 (SD 6.4), respectively, with the difference between the mean number of scheduled exams after the digital awareness campaigns and the number of exams scheduled before and after the digital awareness campaigns being statistically significant (P<.001). CONCLUSIONS: Implementation of the LCS digital awareness campaigns was associated with increased visits to institutional educational Web pages and scheduled LDCT exams. Digital platforms are an important tool to enhance health promotion activities and engagement with patients and providers.

A Novel Model for Predicting Rehospitalization Risk Incorporating Physical Function, Cognitive Status, and Psychosocial Support Using Natural Language Processing.

BACKGROUND: With the increasing focus on reducing hospital readmissions in the United States, numerous readmissions risk prediction models have been proposed, mostly developed through analyses of structured data fields in electronic medical records and administrative databases. Three areas that may have an impact on readmission but are poorly captured using structured data sources are patients' physical function, cognitive status, and psychosocial environment and support. OBJECTIVE OF THE STUDY: The objective of the study was to build a discriminative model using information germane to these 3 areas to identify hospitalized patients' risk for 30-day all cause readmissions. RESEARCH DESIGN: We conducted clinician focus groups to identify language used in the clinical record regarding these 3 areas. We then created a dataset including 30,000 inpatients, 10,000 from each of 3 hospitals, and searched those records for the focus group-derived language using natural language processing. A 30-day readmission prediction model was developed on 75% of the dataset and validated on the other 25% and also on hospital specific subsets. RESULTS: Focus group language was aggregated into 35 variables. The final model had 16 variables, a validated C-statistic of 0.74, and was well calibrated. Subset validation of the model by hospital yielded C-statistics of 0.70-0.75. CONCLUSIONS: Deriving a 30-day readmission risk prediction model through identification of physical, cognitive, and psychosocial issues using natural language processing yielded a model that performs similarly to the better performing models previously published with the added advantage of being based on clinically relevant factors and also automated and scalable. Because of the clinical relevance of the variables in the model, future research may be able to test if targeting interventions to identified risks results in reductions in readmissions.

JOURNAL CLUB: Structured Feedback From Patients on Actual Radiology Reports: A Novel Approach to Improve Reporting Practices.

OBJECTIVE: Patient perceptions of radiology reports are largely unknown. The objective of the present study is to describe our experience receiving structured feedback from patients on actual radiology reports as a means of improving reporting practices. MATERIALS AND METHODS: Eight reports (two for radiographs, two for ultrasound images, two for CT scans, and two for MR images) were randomly selected from our system for review. For each report, patients were asked to rate their level of comprehension, identify any problems in the report, and, in the free-text portion of the feedback form, indicate any questions about the report that they may have. Potentially confounding factors were also examined. RESULTS: A total of 104 patients (46 men and 58 women) participated in the study (for a total of 832 evaluations). The median score for report comprehension was 2.5 (on a scale of 1-5), with the most common problems affecting comprehension identified as "unclear or technical language" (mentioned in 59.6% of evaluations) and the report being "too long" (mentioned in 10.2% of evaluations). A request for an explanation of the report in lay terms (noted in 20.1% of evaluations) was the most common request mentioned in the free-text portion of the feedback form. An inverse relationship existed between report length and patient comprehension (p < 0.001). Patients who had prior experience with their own radiology reports indicated having greater comprehension than did patients with no prior experience (p = 0.003). No correlation between the educational status and report comprehension of the patients was identified (p = 0.488). CONCLUSION: Radiology reports are not well understood by patients, who identify technical language and the long length of reports as the most common problems affecting their comprehension. Longer reports tend to be less well understood.

Fully Automated Deep Learning System for Bone Age Assessment.

Skeletal maturity progresses through discrete phases, a fact that is used routinely in pediatrics where bone age assessments (BAAs) are compared to chronological age in the evaluation of endocrine and metabolic disorders. While central to many disease evaluations, little has changed to improve the tedious process since its introduction in 1950. In this study, we propose a fully automated deep learning pipeline to segment a region of interest, standardize and preprocess input radiographs, and perform BAA. Our models use an ImageNet pretrained, fine-tuned convolutional neural network (CNN) to achieve 57.32 and 61.40% accuracies for the female and male cohorts on our held-out test images. Female test radiographs were assigned a BAA within 1 year 90.39% and within 2 years 98.11% of the time. Male test radiographs were assigned 94.18% within 1 year and 99.00% within 2 years. Using the input occlusion method, attention maps were created which reveal what features the trained model uses to perform BAA. These correspond to what human experts look at when manually performing BAA. Finally, the fully automated BAA system was deployed in the clinical environment as a decision supporting system for more accurate and efficient BAAs at much faster interpretation time (<2 s) than the conventional method.

Hospital Evaluations by Social Media: A Comparative Analysis of Facebook Ratings among Performance Outliers.

BACKGROUND: An increasing number of hospitals and health systems utilize social media to allow users to provide feedback and ratings. The correlation between ratings on social media and more conventional hospital quality metrics remains largely unclear, raising concern that healthcare consumers may make decisions on inaccurate or inappropriate information regarding quality. OBJECTIVES: The purpose of this study was to examine the extent to which hospitals utilize social media and whether user-generated metrics on Facebook(®) correlate with a Hospital Compare(®) metric, specifically 30-day all cause unplanned hospital readmission rates. DESIGN AND PARTICIPANTS: This was a retrospective cross-sectional study conducted among all U.S. hospitals performing outside the confidence interval for the national average on 30-day hospital readmission rates as reported on Hospital Compare. Participants were 315 hospitals performing better than U.S. national rate on 30-day readmissions and 364 hospitals performing worse than the U.S. national rate. MAIN MEASURES: The study analyzed ratings of hospitals on Facebook's five-star rating scale, 30-day readmission rates, and hospital characteristics including beds, teaching status, urban vs. rural location, and ownership type. KEY RESULTS: Hospitals performing better than the national average on 30-day readmissions were more likely to use Facebook than lower-performing hospitals (93.3 % vs. 83.5 %; p < 0.01). The average rating for hospitals with low readmission rates (4.15 ± 0.31) was higher than that for hospitals with higher readmission rates (4.05 ± 0.41, p < 0.01). Major teaching hospitals were 14.3 times more likely to be in the high readmission rate group. A one-star increase in Facebook rating was associated with increased odds of the hospital belonging to the low readmission rate group by a factor of 5.0 (CI: 2.6-10.3, p < 0.01), when controlling for hospital characteristics and Facebook-related variables. CONCLUSIONS: Hospitals with lower rates of 30-day hospital-wide unplanned readmissions have higher ratings on Facebook than hospitals with higher readmission rates. These findings add strength to the concept that aggregate measures of patient satisfaction on social media correlate with more traditionally accepted measures of hospital quality.

Creating a Patient-Centered Radiology Practice Through the Establishment of a Diagnostic Radiology Consultation Clinic.

OBJECTIVE: The purposes of this study were to assess the feasibility of and to create a referral mechanism for a diagnostic radiology consultation clinic. SUBJECTS AND METHODS: A pilot program was instituted with patients from a single primary care clinic over a 3-week period. Patients with findings of common problems at routine imaging, such as atherosclerosis, emphysema, and hepatic steatosis, were eligible to participate. As the patients arrived for their routine primary care visits, office staff informed them of the opportunity to formally meet with a radiologist to review their most recent imaging findings. The office staff of the primary care clinic then contacted the radiologist covering the diagnostic radiology consultation clinic to schedule a consultation. A survey was administered before and after the session. RESULTS: Twenty-two patients participated (88% participation rate). Participants rated the consultation as very helpful (mean, 4.8 on 1-5 scale), and all participants would take the opportunity to review studies with the radiologist again. Significantly more patients preferred the involvement of the radiologist in communicating the results of an imaging examination after the consultation compared with before the consultation (p = 0.001). After the consultation session, patients had significantly improved understanding of a radiologist's role (p = 0.004), and all participants were able to correctly identify the radiologist as a physician who interprets medical images. CONCLUSION: A referral mechanism for a diagnostic radiology consultation clinic can be effectively integrated into the everyday workflow of both the referring physician and the radiologist. The consultations are useful to patients and help to increase their awareness of the role of the radiologist.

Rethinking the role of the radiologist: enhancing visibility through both traditional and nontraditional reporting practices.

The written radiology report is the primary method by which the radiologist communicates examination findings to the referring physician and the patient. Unfortunately, despite recent efforts to improve radiology reporting practices, dissatisfaction with regard to traditional reporting practices continues among referring physicians and patients. Moreover, the current reporting paradigm limits the amount of interaction that radiologists can have with patients and referring physicians. To address these issues, organizational efforts have been made by the Radiological Society of North America and the American College of Radiology to educate the public about the role of the radiologist in patient care, provide resources for individuals to learn more about the practice of radiology, and assist radiologists in creating a more "patient-centered" practice. In addition, individual radiologists may want to consider making adjustments to their reporting practices to improve communication of examination findings and increase patient awareness of the radiologist's role in clinical care. Many opportunities exist for radiologists to make such changes, both within the traditional reporting model (decreasing barriers to communication, educating patients) and with implementation of various nontraditional reporting practices (eg, delivering examination results directly to patients, receiving structured feedback on reports, establishing a diagnostic radiology consultation clinic). The authors provide guidance for radiologists who wish to enhance their visibility in clinical care with use of both traditional and nontraditional reporting practices.

Assessment of Filtered Back Projection, Adaptive Statistical, and Model-Based Iterative Reconstruction for Reduced Dose Abdominal Computed Tomography.

PURPOSE: To compare standard of care and reduced dose (RD) abdominal computed tomography (CT) images reconstructed with filtered back projection (FBP), adaptive statistical iterative reconstruction (ASIR), model-based iterative reconstruction (MBIR) techniques. MATERIALS AND METHODS: In an Institutional Review Board-approved, prospective clinical study, 28 patients (mean age 59 ± 13 years ), undergoing clinically indicated routine abdominal CT on a 64-channel multi-detector CT scanner, gave written informed consent for acquisition of an additional RD (<1 milli-Sievert) abdomen CT series. Sinogram data of RD series were reconstructed with FBP, ASIR, and MBIR and compared with FBP images of standard dose abdomen CT. Two radiologists performed randomized, independent, and blinded comparison for lesion detection, lesion margin, visibility of normal structures, and diagnostic confidence. RESULTS: Mean CT dose index volume was 10 ± 3.4 mGy and 1.3 ± 0.3 mGy for standard and RD CT, respectively. There were 73 "true positive" lesions detected on standard of care CT. Nine lesions (<8 mm in size) were missed on RD abdominal CT images which included liver lesions, liver cysts, kidney cysts, and paracolonic abscess. These lesions were missed regardless of patient size and types of iterative reconstruction techniques used for reconstruction of RD data sets. The visibility of lesion margin was suboptimal in (23/28) patients with RD FBP, (15/28) patients with RD ASIR, and (14/28) patients with RD MBIR compared to standard of care FBP images (P < 0.001). Diagnostic confidence for the assessment of lesions on RD images was suboptimal in most patients regardless of iterative reconstruction techniques. CONCLUSIONS: Clinically significant lesions (< 8 mm) can be missed on abdominal CT examinations acquired at a CT dose index volume of 1.3 mGy regardless of patients' size and reconstruction techniques (FBP, ASIR, and MBIR).

Coronary artery calcification is often not reported in pulmonary CT angiography in patients with suspected pulmonary embolism: an opportunity to improve diagnosis of acute coronary syndrome.

OBJECTIVE: In patients with suspected pulmonary thromboembolism (PTE), coronary artery calcification (CAC) can be an incidental finding on pulmonary CT angiography. We evaluated the frequency of CAC not being reported and its association with a diagnosis of acute coronary syndrome (ACS). MATERIALS AND METHODS: Data of 469 consecutive patients who were referred to the emergency radiology department for pulmonary CT angiography of suspected PTE were reviewed. Radiology reports were rechecked and positive CAC findings were recorded. All pulmonary CT angiograms were reevaluated by one radiologist and CAC findings were recorded. The rates of ACS and PTE as final diagnoses for that hospital admission were calculated. The association between CAC and ACS diagnosis was assessed in different subgroups of patients. RESULTS: Approximately 11.1% of patients had PTE and 43.8% had CAC. The incidence of CAC was significantly higher in patients with an ACS diagnosis than in those without ACS (56.2% vs 40.4%, respectively; odds ratio [OR] = 1.9). There was a strong positive association (OR = 3.5) between CAC and ACS in younger patients (men ≤ 45 years and women ≤ 55 years), patients without PTE (OR = 2.2), and those without cardiometabolic risk factors (OR = 3.8). CAC was not reported in 45% of patients (n = 98) with positive CAC findings on imaging. ACS was the final diagnosis in 31.6% of patients with unreported CAC. There was a significant association between CAC and ACS in patients with unreported CAC (OR = 2.2). This association was more prominent in the subgroups described. CONCLUSION: CAC is often not reported in pulmonary CT angiography studies. CAC is a significant predictor of ACS particularly in younger patients, patients without PTE, and those without cardiometabolic risk factors. Especially in these subgroups, radiologists should assess CAC findings.

Radiologists' role in the communication of imaging examination results to patients: perceptions and preferences of patients.

OBJECTIVE: It has been suggested that radiology reporting practices would be improved if radiologists were to discuss the results of an examination directly with the patient. The attitudes and preferences of patients with regard to direct communication with the radiologist are not well-defined. The purpose of this study was to survey patients about their preferred method of receiving radiologic results. MATERIALS AND METHODS: An anonymous survey was distributed to adult patients undergoing contrast-enhanced CT or MRI over a 2-week period in June 2013. RESULTS: The response rate was 58.4% (642 responses). For normal examination results, the preferred mode of communication was a telephone call from the ordering physician (34.1%); only 12% of respondents opted for a telephone call from the radiologist, and 2.6% chose a face-to-face meeting with the radiologist. For abnormal test results, the preferred mode of communication was also a telephone call from the ordering physician (49.8%); 14.4% of respondents selected a telephone call from the radiologist, and 8.3% chose a face-to-face meeting with the radiologist. Patients preferred receiving very detailed versions of radiology reports for both normal (46.4%) and abnormal (81.8%) test results. Patients also expressed a desire to have access to at least key images from their examinations. CONCLUSION: Patients prefer receiving both normal and abnormal examination results from the physicians who ordered the examination rather than the radiologist. They also prefer to receive very detailed examination results rather than a brief summary in lay terms.

Effect of localizer radiograph on radiation dose associated with automatic exposure control: human cadaver and patient study.

PURPOSE: To evaluate the effect of localizing radiograph on computed tomography (CT) radiation dose associated with automatic exposure control with a human cadaver and patient study. MATERIALS AND METHODS: Institutional review board approved the study with a waiver of informed consent. Two chest CT image series with fixed tube current and combined longitudinal-angular automatic exposure control (AEC) were acquired in a human cadaver (64-year-old man) after each of the 8 combinations of localizer radiographs (anteroposterior [AP], AP lateral, AP-posteroanterior [PA], lateral AP, lateral PA, PA, PA-AP, and PA lateral). Applied effective milliampere second, volume CT dose index (CTDIvol) and image noise were recorded for all 24-image series. Volume CT dose indexes were also recorded in 20 patients undergoing chest and abdominal CT after PA and PA-lateral radiographs with the use of AEC. Data were analyzed using analysis of variance and linear correlation tests. RESULTS: With AEC, the CTDIvol fluctuates with the number and projection of localizer radiographs (P < 0.0001). Lowest CTDIvol values are seen when 2 orthogonal localizer radiographs are acquired, whereas highest values are seen when single PA or AP-PA projection localizer radiographs are acquired for planning (P < 0.0001). In 20 patients, CT scanning with AEC after acquisition of 2 orthogonal projection localizer radiographs was associated with significant reduction in radiation dose compared to PA projection radiographs alone (P < 0.0001). CONCLUSIONS: When scanning with AEC, acquisition of 2 orthogonal localizer radiographs is associated with lower CTDIvol compared to a single localizer radiograph.

Structured feedback from referring physicians: a novel approach to quality improvement in radiology reporting.

OBJECTIVE: There is considerable interest in improving radiology reporting practices through peer review, and the inclusion of structured feedback from referring physicians may improve this process. The purpose of this article is to evaluate the feasibility of this type of novel peer-review system. MATERIALS AND METHODS: Five referring physicians from our institution participated as reviewers. Reports from abdominal CT, chest CT, brain MRI, and abdominal ultrasound were eligible for review if the indication was "abdominal pain," "shortness of breath," "headache," and "pain," respectively. Reports were excluded if the examinations were normal or ordered as follow-up. Forty-eight reports (12 from each group) were then selected at random and distributed to the referring physicians along with the clinical scenario and an evaluation form for each examination. RESULTS: The reports were found to be clinically useful (average, 3.8 on a 1-5 scale), allowing for good confidence in clinical decision making (average, 3.7). The most common problems were unclear language, typographical errors, and reports that did not answer the clinical question. Of the reports, 35.4% contained recommendations for further diagnosis or treatment, and 84.7% of these recommendations were deemed clinically appropriate. The participating physicians thought that the results of 31.2% of the examinations should have been directly communicated to the ordering provider. CONCLUSION: Radiology reports and recommendations were clinically useful, even though problems with language, typographical errors, answering the clinical question, and direct communication of examination results were identified. Structured feedback from referring physicians is a novel approach to the peer-review process that may identify problems that go unnoticed by the radiologist, thereby improving reporting practices.

Pertinent reportable incidental cardiac findings on chest CT without electrocardiography gating: review of 268 consecutive cases.

BACKGROUND: Pertinent reportable cardiac findings on non-electrocardiography (ECG)-gated chest CT examinations have become easier to detect given recent advancements in multidetector CT technology. However, those findings are easily overlooked on routine chest CT without ECG gating given residual inherent cardiac motion artifact and non-cardiac indications. PURPOSE: To describe and quantify the types of pertinent reportable cardiac findings that can be detected on chest CT examinations without ECG gating and evaluate how often they were reported. MATERIAL AND METHODS: Two radiologists retrospectively reviewed (blinded to the original interpretation) 268 consecutive routine adult chest CT examinations without ECG gating for the presence of pertinent reportable cardiac findings. Retrospective interpretations were then compared with the original radiological reports. RESULTS: One hundred and sixty-three patients (61%) had pertinent reportable cardiac findings. The findings encountered included: coronary artery disease (n = 131; 80.0%), coronary artery bypass grafts (n = 10; 6.1%), left ventricular aneurysm (n = 1; 0.6%), valve calcification (n = 131; 80.0%), valve repair/replacement (n = 5; 3.1%), pericardial effusion (n = 33; 20.2%), left atrial appendage thrombus (n = 1; 0.6%), cardiac mass (n = 1; 0.6%), and cardiac chamber enlargement (n = 29; 17.8%). On the original radiological reports 22.3% of the pertinent reportable cardiac findings, detected by the two radiologists retrospectively, were not reported. CONCLUSION: Detection of pertinent reportable cardiac findings on routine chest CT examinations without ECG gating is possible. The high volume of chest CT examinations without ECG gating represents an opportunity for radiologists to comment on the presence or absence of cardiac disease which may influence future clinical decisions.

Oxygen saturation can predict pediatric pneumonia in a resource-limited setting.

BACKGROUND: The World Health Organization (WHO) recommends using age-specific respiratory rates for diagnosing pneumonia in children. Past studies have evaluated the WHO criteria with mixed results. OBJECTIVE: We examined the accuracy of clinical and laboratory factors for diagnosing pediatric pneumonia in resource-limited settings. METHODS: We conducted a retrospective chart review of children under 5 years of age presenting with respiratory complaints to three rural hospitals in Rwanda who had received a chest radiograph. Data were collected on the presence or absence of 31 historical, clinical, and laboratory signs. Chest radiographs were interpreted by pediatric radiologists as the gold standard for diagnosing pneumonia. Overall correlation and test characteristics were calculated for each categorical variable as compared to the gold standard. For continuous variables, we created receiver operating characteristic (ROC) curves to determine their accuracy for predicting pneumonia. RESULTS: Between May 2011 and April 2012, data were collected from 147 charts of children with respiratory complaints. Approximately 58% of our sample had radiologist-diagnosed pneumonia. Of the categorical variables, a negative blood smear for malaria (χ(2) = 6.21, p = 0.013) and the absence of history of asthma (χ(2) = 4.48, p = 0.034) were statistically associated with pneumonia. Of the continuous variables, only oxygen saturation had a statistically significant area under the ROC curve (AUC) of 0.675 (95% confidence interval [CI] 0.581-0.769 and p = 0.001). Respiratory rate had an AUC of 0.528 (95% CI 0.428-0.627 and p = 0.588). CONCLUSION: Oxygen saturation was the best clinical predictor for pediatric pneumonia and should be further studied in a prospective sample of children with respiratory symptoms in a resource-limited setting.

Cardiac Incidental Findings on Abdominopelvic Computed Tomography: Prevalence and Association with Subsequent Cardiovascular Events.

PURPOSE: The aim of this study was to assess the prevalence of reportable cardiac findings detected on abdominopelvic CTs and the association with subsequent cardiovascular events. MATERIALS AND METHODS: We performed a retrospective search of electronic medical record of patients who underwent abdominopelvic CT between November 2006 and November 2011 with a clinical history of upper abdominal pain. A radiologist blinded to the original CT report reviewed all 222 cases for the presence of pertinent reportable cardiac findings. The original CT report was also evaluated for documentation of pertinent reportable cardiac findings. The following findings were recorded on all CTs: presence of coronary calcification, fatty metaplasia, ventricle wall thinning and thickening, valve calcification or prosthesis, heart/chamber enlargement, aneurysm, mass, thrombus, device, air within ventricles, abnormal pericardium, prior sternotomy, and adhesions if prior sternotomy. Medical records were reviewed to identify cardiovascular events on follow-up in patients with the presence or absence of cardiac findings. We compared the distribution findings in patients with and without cardiac events using the Wilcoxon test (for continuous variables) and the Pearson's chi-squared test (for categorical variables). RESULTS: Eighty-five of 222 (38.3%) patients (52.7% females, median age 52.5 years) had at least one pertinent reportable cardiac finding on the abdominopelvic CT, with a total of 140 findings in this group. From the total 140 findings, 100 (71.4%) were not reported. The most common findings seen on abdominal CTs were: coronary artery calcification (66 patients), heart or chamber enlargement (25), valve abnormality (19), sternotomy and surgery signs (9), LV wall thickening (7), device (5), LV wall thinning (2), pericardial effusion (5), and others (3). After a mean follow-up of 43.9 months, 19 cardiovascular events were found in the cohort (transient ischemic attack, cerebrovascular accident, myocardial infarction, cardiac arrest, acute arrhythmia, palpitation, syncope and acute chest pain). Only 1 event occurred in the group of patients with no incidental pertinent reportable cardiac findings (1/137 = 0.73%). All other 18 events occurred in patients with incidental pertinent reportable cardiac findings (18/85 = 21.2%), which was significantly different (p < 0.0001). One out of the total 19 events in the overall group (5.24%) occurred in a patient with no incidental pertinent reportable cardiac findings while 18 of 19 total events (94.74%) occurred with patients with incidental pertinent reportable cardiac findings, which was also significantly different (p < 0.001). Fifteen of the total events (79%) occurred in patients in whom the incidental pertinent reportable cardiac findings were not reported, which was significantly different (p < 0.001) from the four events that occurred in patients in whom the incidental pertinent reportable cardiac findings were reported or had no findings. CONCLUSIONS: Incidental pertinent reportable cardiac findings are common on abdominal CTs and are frequently not reported by radiologists. These findings are of clinical relevance since patients with pertinent reportable cardiac findings have a significantly higher incidence of cardiovascular events on follow-up.

Radiation dose reduction with Sinogram Affirmed Iterative Reconstruction technique for abdominal computed tomography.

PURPOSE: The objective of this study was to assess the effect of Sinogram Affirmed Iterative Reconstruction (SAFIRE) and filtered back-projection (FBP) techniques on abdominal computed tomography (CT) performed with 50% and 75% radiation dose reductions. METHODS: Twenty-four patients (mean age, 64 ± 14 years; male-female ratio, 10:14) gave informed consent for an institutional review board-approved prospective study involving acquisition of additional research images through the abdomen on 128-slice multi-detector-row CT (SOMATOM Definition Flash) at quality reference mAs of 100 (50% lower dose) and 50 (75% lower dose) over a scan length of 10 cm using combined modulation (CARE Dose 4D). Standard-of-care abdominal CT was performed at 200 quality reference mAs, with remaining parameters held constant. The 50- and 100-mAs data sets were reconstructed with FBP and at 4 SAFIRE settings (S1, S2, S3, S4). Higher number of SAFIRE settings denotes increased strength of the algorithm resulting in lower image noise. Two abdominal radiologists independently compared the FBP and SAFIRE images for lesion number, location, size and conspicuity, and visibility of small structures, image noise, and diagnostic confidence. Objective noise and Hounsfield units (HU) were measured in the liver and the descending aorta. RESULTS: All 43 lesions were detected on both FBP and SAFIRE images. Minor blocky, pixelated appearance of 50% and 75% reduced dose images was noted at S3 and S4 SAFIRE but not at S1 and S2 settings. Subjective noise was suboptimal in both 50% and 75% lower-dose FBP images but was deemed acceptable on all SAFIRE settings. Sinogram Affirmed Iterative Reconstruction images were deemed acceptable in all patients at 50% lower dose and in 22 of 24 patients at 75% lower dose. As compared with 75% reduced dose FBP, objective noise was lower by 22.8% (22.9/29.7), 35% (19.3/29.7), 44.3% (16.7/29.3), and 54.8% (13.4/29.7) on S1 to S4 settings, respectively (P < 0.001). CONCLUSIONS: Sinogram Affirmed Iterative Reconstruction-enabled reconstruction provides abdominal CT images without loss in diagnostic value at 50% reduced dose and in some patients also at 75% reduced dose.