Low-dose CT examination for rib fracture evaluation: A pilot study

Diagnostic Accuracy of Ultra-Low Dose CT Compared to Standard Dose CT for Identification of Fresh Rib Fractures by Deep Learning Algorithm

The present study aimed to evaluate the diagnostic accuracy of ultra-low dose computed tomography (ULD-CT) compared to standard dose computed tomography (SD-CT) in discerning recent rib fractures using a deep learning algorithm detection of rib fractures (DLADRF). A total of 158 patients undergoing forensic diagnosis for rib fractures were included in this study: 50 underwent SD-CT, and 108 were assessed using ULD-CT. Junior and senior radiologists independently evaluated the images to identify and characterize the rib fractures. The sensitivity of rib fracture diagnosis by radiologists and radiologist + DLADRF was better using SD-CT than ULD-CT. However, the diagnosis sensitivity of DLADRF using ULD-CT alone was slightly more than SD-CT. Nonetheless, no substantial differences were observed in specificity, positive predictive value, and negative predictive value between SD-CT and ULD-CT by the same radiologist, radiologist + DLADRF, and DLADRF (P > 0.05). The area under the curve (AUC) of receiver operating characteristic indicated that senior radiologist + DLADRF was significantly better than senior and junior radiologists, junior radiologists + DLADRF, and DLADRF alone using SD-CT or ULD-CT (all P < 0.05). Also, junior radiologists + DLADRF was better with ULD-CT than senior and junior radiologists (P < 0.05). The AUC of the rib fracture diagnosed by senior radiologists did not differ from DLADRF using ULD-CT. Also, no significant differences were observed between junior + AI and senior and between junior and DLADRF using SD-CT. DLADRF enhanced the diagnostic performance of radiologists in detecting recent rib fractures. The diagnostic outcomes between SD-CT and ULD-CT across radiologists' experience and DLADRF did not differ significantly.

Feasibility of rib fracture detection in low-dose computed tomography images with a large, multicenter datasets-based model

Purpose

To evaluate the feasibility of rib fracture detection in low-dose computed tomography (CT) images with a RetinaNet-based approach and to evaluate the potential of lowdose CT for rib fracture detection compared with regular-dose CT images.

Materials and methods

The RetinaNet-based deep learning model was trained using 7300 scans with 50,410 rib fractures that were used as internal training datasets from four multicenter. The external test datasets consisted of both regular-dose and low-dose chest-abdomen CT images of rib fractures; the MICCAI 2020 RibFrac Challenge Dataset was used as the public dataset. Radiologists' interpretations were used as reference standards. The performance of the model in rib fracture detection was compared with the radiologists' interpretation.

Results

In total, 728 traumatic rib fractures of 100 patients [60 men (60 %); mean age, 53.45 ± 11.19 (standard deviation (SD)); range, 18-77 years] were assessed in the external datasets. In these patients, the regular-dose group had a mean CT dose index volume (CTDIvol) of 7.18 mGy (SD: 2.22) and a mean dose length product (DLP) of 305.38 mGy cm (SD: 95.31); the low-dose group had a mean CTDIvol of 2.79 mGy (SD: 1.11) and a mean DLP of 131.52 mGy cm (SD: 55.58). The sensitivity of the RetinaNet-based model and that of the radiologists was 0.859 and 0.721 in the low-dose CT images and 0.886 and 0.794 in the regular-dose CT images, respectively.

Conclusions

These findings indicate that the RetinaNet-based model can detect rib fractures in low-dose CT images with a robust performance, indicating its feasibility in assisting radiologists with rib fracture diagnosis.

Value of Thoracic CT in Blunt Trauma Patients With High Glasgow Coma Scale and Low Injury Severity Scale

OBJECTIVE

To determine the value of ordering a routine chest CT (CCT) in patients with blunt trauma presenting to the emergency department with a high GCS and low ISS, we retrospectively collected patient data including CT scan results, when physical examination and initial chest X-ray were normal in the trauma bay area.

METHODS

A retrospective data collection of 901 consecutive blunt trauma patients seen in the ED between 2017 and 2019 was analyzed. Data included physical examination, age, gender, current use of anticoagulation therapy, comorbid conditions, as well as the result of radiologic images, hospital length of stay, surgical intervention, and mortality. The patients were divided into two groups: group one (patients with negative physical examination; chest x-ray and CT) and group 2 (negative physical examination, positive or negative chest x-ray, and positive CT). Statistical analysis was performed using student's t-test and chi-square test.

RESULTS

Of the 901 patients there were 489 (54%) males and 412 (46%) females with a mean age of 56 years. There were 461 patients who had a physical examination, chest x-ray, abdominal and CCT done. Group one included 442 (96%) patients, with negative physical examination, negative chest X-ray and CT scan. In group 2, 19 (4%) patients who had positive CT and or chest x-ray. Both groups were similar in GCS and ISS. Of the 19 patients, sixteen patients had a positive CCT, and thirteen of those had a positive chest x-ray. In the three patients who had negative physical examination and chest x-ray, the CT findings included one with a nondisplaced 10th rib fracture and two patients with osteoporotic compression fractures of dorsal vertebrae. The rate of both chest x-ray and CCT being positive among a group of screened patients was 16% (3/19) and the rate of a negative chest x-ray but positive CT was 16% (3/19). The odds ratio between the two outcomes was one.

CONCLUSION

In blunt trauma patients presenting to the ED with a high GCS and low ISS score, when initial physical examination and chest x-ray are negative, routine CCT is of little value.

The value of deep learning-based computer aided diagnostic system in improving diagnostic performance of rib fractures in acute blunt trauma

BACKGROUND

To evaluate the value of a deep learning-based computer-aided diagnostic system (DL-CAD) in improving the diagnostic performance of acute rib fractures in patients with chest trauma.

MATERIALS AND METHODS

CT images of 214 patients with acute blunt chest trauma were retrospectively analyzed by two interns and two attending radiologists independently firstly and then with the assistance of a DL-CAD one month later, in a blinded and randomized manner. The consensusdiagnosis of fib fracture by another two senior thoracic radiologists was regarded as reference standard. The rib fracture diagnostic sensitivity, specificity, positive predictive value, diagnostic confidence and mean reading time with and without DL-CAD were calculated and compared.

RESULTS

There were 680 rib fracture lesions confirmed as reference standard among all patients. The diagnostic sensitivity and positive predictive value of interns weresignificantly improved from (68.82%, 84.50%) to (91.76%, 93.17%) with the assistance of DL-CAD, respectively. Diagnostic sensitivity and positive predictive value of attendings aided by DL-CAD (94.56%, 95.67%) or not aided (86.47%, 93.83%), respectively. In addition, when radiologists were assisted by DL-CAD, the mean reading time was significantly reduced, and diagnostic confidence was significantly enhanced.

CONCLUSIONS

DL-CAD improves the diagnostic performance of acute rib fracture in chest trauma patients, which increases the diagnostic confidence, sensitivity, and positive predictive value for radiologists. DL-CAD can advance the diagnostic consistency of radiologists with different experiences.

Computed tomography with low-dose radiation versus standard-dose radiation for diagnosing fractures: systematic review and meta-analysis

BACKGROUND

Computed tomography (CT) accounts for 13% of all radiological examinations in the United States and 40-70% of the radiation that patients receive. Even with the advent of magnetic resonance imaging (MRI), CT continues to be the gold standard for diagnosing bone fractures. There is uncertainty as to whether CT with a low radiation dose has a fracture detection rate similar to that of standard-dose CT.

OBJECTIVE

To determine the detection rate of low-dose radiation CT and standard-dose radiation CT for fractures, in patients with suspected fractures.

DESIGN AND SETTING

Systematic review of comparative studies on diagnostic accuracy within the evidence-based health program at a federal university in São Paulo (SP), Brazil.

METHODS

We searched the electronic databases Cochrane Library, MEDLINE, EMBASE and LILACS up to June 29, 2020, for studies evaluating the detection rates of low-dose CT and standard-dose CT for diagnosing bone fractures. The Research Triangle Institute (RTI) item bank tool was used for methodological quality evaluation.

RESULTS

The fracture detection rate according to the number of bones evaluated, using CT with low-dose radiation was 20.3%, while with standard-dose radiation it was 19.2%, and the difference between the methods was not significant. The fracture detection rate according to the number of patients, using CT with low-dose radiation was 56.0%, while with standard-dose radiation it was 58.7%, and this difference between the methods was not significant, either.

CONCLUSION

CT with low-dose radiation presented detection rates similar to those of CT with standard-dose radiation, regardless of the bones evaluated.

REGISTRATION NUMBER

CRD42019148491 at the PROSPERO database.

Magnetic Resonance Imaging has Better Accuracy in Detecting New-Onset Rib Fractures as Compared to Computed Tomography

Background

The aim of this study was to explore the magnetic resonance imaging (MRI) manifestations of new-onset rib fractures and determine the utility of MRI through a comparative study of MRI and computed tomography (CT).

Material/Methods

Twenty-one patients with mild chest trauma who sought medical care from February 2019 to June 2020 were included in this study. The patients were subjected to CT and MRI scanning within 2 weeks after the trauma, and CT rescanning was scheduled 4–8 weeks later to classify rib fractures and determine the diagnostic accuracy of MRI and CT for new-onset rib fractures.

Results

Seventy-seven rib fractures were confirmed by CT rescanning, of which 72 (93.51%) were type I fractures and 5 (6.49%) were type II. MRI identified 76 fractures, of which 3 were false positive, with the diagnostic accuracy rate of 91.25% and sensitivity rate of 94.81%. Among them, type I fractures (n=71, 3 were false positive) showed the MRI “sandwich” sign (heterogeneous high-signal shadow within bone marrow of the inner layer, low-signal bony cortex of the middle layer, and high-signal subperiosteal effusion of the outer layer) in T2-weighted fat-suppressed sequences; type II fractures (n=5) displayed intramedullary high-signal intensities and no subperiosteal effusion. Forty-four fractures (all type I) were discovered in the initial CT examination, and the corresponding diagnostic accuracy rate and sensitivity rate were 57.14%, which were lower than that of MRI.

Conclusions

MRI is highly sensitive for new-onset rib fractures, especially type I, so it is a preferred method for patients with mild chest trauma.

Deep-learning-assisted detection and segmentation of rib fractures from CT scans: Development and validation of FracNet

BACKGROUND

Diagnosis of rib fractures plays an important role in identifying trauma severity. However, quickly and precisely identifying the rib fractures in a large number of CT images with increasing number of patients is a tough task, which is also subject to the qualification of radiologist. We aim at a clinically applicable automatic system for rib fracture detection and segmentation from CT scans.

METHODS

A total of 7,473 annotated traumatic rib fractures from 900 patients in a single center were enrolled into our dataset, named RibFrac Dataset, which were annotated with a human-in-the-loop labeling procedure. We developed a deep learning model, named FracNet, to detect and segment rib fractures. 720, 60 and 120 patients were randomly split as training cohort, tuning cohort and test cohort, respectively. Free-Response ROC (FROC) analysis was used to evaluate the sensitivity and false positives of the detection performance, and Intersection-over-Union (IoU) and Dice Coefficient (Dice) were used to evaluate the segmentation performance of predicted rib fractures. Observer studies, including independent human-only study and human-collaboration study, were used to benchmark the FracNet with human performance and evaluate its clinical applicability. A annotated subset of RibFrac Dataset, including 420 for training, 60 for tuning and 120 for test, as well as our code for model training and evaluation, was open to research community to facilitate both clinical and engineering research.

FINDINGS

Our method achieved a detection sensitivity of 92.9% with 5.27 false positives per scan and a segmentation Dice of 71.5%on the test cohort. Human experts achieved much lower false positives per scan, while underperforming the deep neural networks in terms of detection sensitivities with longer time in diagnosis. With human-computer collobration, human experts achieved higher detection sensitivities than human-only or computer-only diagnosis.

INTERPRETATION

The proposed FracNet provided increasing detection sensitivity of rib fractures with significantly decreased clinical time consumed, which established a clinically applicable method to assist the radiologist in clinical practice.

FUNDING

A full list of funding bodies that contributed to this study can be found in the Acknowledgements section. The funding sources played no role in the study design; collection, analysis, and interpretation of data; writing of the report; or decision to submit the article for publication .