Correlation between early direct communication of positive CT pulmonary angiography findings and improved clinical outcomes

Automatic detection of pulmonary embolism on computed tomography pulmonary angiogram scan using a three-dimensional convolutional neural network

OBJECTIVE

To propose a convolutional neural network (EmbNet) for automatic pulmonary embolism detection on computed tomography pulmonary angiogram (CTPA) scans and to assess its diagnostic performance.

METHODS

305 consecutive CTPA scans between January 2019 and December 2021 were enrolled in this study (142 for training, 163 for internal validation), and 250 CTPA scans from a public dataset were used for external validation. The framework comprised a preprocessing step to segment the pulmonary vessels and the EmbNet to detect emboli. Emboli were divided into three location-based subgroups for detailed evaluation: central arteries, lobar branches, and peripheral regions. Ground truth was established by three radiologists.

RESULTS

The EmbNet's per-scan level sensitivity, specificity, positive predictive value (PPV), and negative predictive value were 90.9%, 75.4%, 48.4%, and 97.0% (internal validation) and 88.0%, 70.5%, 42.7%, and 95.9% (external validation). At the per-embolus level, the overall sensitivity and PPV of the EmbNet were 86.0% and 61.3% (internal validation), and 83.5% and 57.5% (external validation). The sensitivity and PPV of central emboli were 89.7% and 52.0% (internal validation), and 94.4% and 43.0% (external validation); of lobar emboli were 95.2% and 76.9% (internal validation), and 93.5% and 72.5% (external validation); and of peripheral emboli were 82.6% and 61.7% (internal validation), and 80.2% and 59.4% (external validation). The average false positive rate was 0.45 false emboli per scan (internal validation) and 0.69 false emboli per scan (external validation).

CONCLUSION

The EmbNet provides high sensitivity across embolus locations, suggesting its potential utility for initial screening in clinical practice.

An Enhanced Mask R-CNN Approach for Pulmonary Embolism Detection and Segmentation

Pulmonary embolism (PE) refers to the occlusion of pulmonary arteries by blood clots, posing a mortality risk of approximately 30%. The detection of pulmonary embolism within segmental arteries presents greater challenges compared with larger arteries and is frequently overlooked. In this study, we developed a computational method to automatically identify pulmonary embolism within segmental arteries using computed tomography (CT) images. The system architecture incorporates an enhanced Mask R-CNN deep neural network trained on PE-containing images. This network accurately localizes pulmonary embolisms in CT images and effectively delineates their boundaries. This study involved creating a local data set and evaluating the model predictions against pulmonary embolisms manually identified by expert radiologists. The sensitivity, specificity, accuracy, Dice coefficient, and Jaccard index values were obtained as 96.2%, 93.4%, 96.%, 0.95, and 0.89, respectively. The enhanced Mask R-CNN model outperformed the traditional Mask R-CNN and U-Net models. This study underscores the influence of Mask R-CNN's loss function on model performance, providing a basis for the potential improvement of Mask R-CNN models for object detection and segmentation tasks in CT images.

Discrepancy rate and clinical impact of preliminary reports from radiology residents

Background

Residents usually cover night and weekend shifts issuing the preliminary reading of radiological studies in university hospitals. This is essential to strengthening decision-making skills when facing complex cases independently. However, there should be a balance between patient safety and academic experience since some concern has been expressed about the accuracy of the interpretations generated by trainees. This work aims to evaluate and characterize the discrepancies in preliminary reports issued by radiology residents.

Material and methods

Radiologists filled out a questionnaire to evaluate preliminary reports of trainees considering diagnosis, findings description, clinical approach changes, and critical findings. Analysis was performed considering modality, imaging type, body part, and resident academic year. A Chi-square test with a significance level α of 0.05 was used to make group comparisons.

Results

A total of 9072 studies were reviewed. Major and minor overall discrepancy rates were 1.7% and 8.3%, respectively. Minor discrepancy rate, findings description, and critical findings identification improved with increasing academic year, both overall and by modality. Discrepancy rates were lower for CT than MR and neuroimaging than for body-imaging studies. The highest major and minor discrepancy rates as abdomen/pelvis CT and lumbar-spine MR, respectively. Two percent of reports presented discrepancies that could generate a medical approach change.

Conclusion

Discrepancy rates are low and comparable with those reported in the literature. These rates tend to improve as the resident's academic year increases. Our results suggest that radiology residents' coverage of night shifts and weekends is a practice that benefits the educational process without negatively impacting patient safety.

Decreased Hospital Length of Stay for ICH and PE after Adoption of an Artificial Intelligence-Augmented Radiological Worklist Triage System

The purpose of the study was to determine whether there was a difference in the length of stay (LOS) for inpatients diagnosed with intracranial hemorrhage (ICH) or pulmonary embolism (PE) prior to and following implementation of an (AI) triage software. A retrospective review was performed for patients that underwent CT imaging procedures related to ICH and PE from April 2016 to October 2019. All patient encounters that included noncontrast head computed tomography (CT) or CT chest angiogram (CTCA) procedures, identified by the DICOM study descriptions, from April 2016 to April 2019 were included for ICH and PE, respectively. All patients that were diagnosed with ICH or PE were identified using ICD9 and ICD10 codes. Three separate control groups were defined as follows: (i) all remaining patients that underwent the designated imaging studies, (ii) patients diagnosed with hip fractures, and (iii) all hospital wide encounters, during the study period. Pre-AI and post-AI time periods were defined around the deployment dates of the ICH and PE modules, respectively. The reduction in LOS was 1.30 days (95% C.I. 0.1–2.5), resulting in an observed percentage decrease of 11.9% (p value = 0.032), for ICH and 2.07 days (95% C.I. 0.1–4.0), resulting in an observed percentage decrease of 26.3% (p value = 0.034), for PE when comparing the pre-AI and post-AI time periods. Reductions in LOS were observed in the ICH pre-AI and post-AI time period group for patients that were not diagnosed with ICH, but that underwent related imaging, 0.46 days (95% C.I. 0.1–0.8) resulting in an observed percentage decrease of 5% (p value = 0.018), and inpatients that were diagnosed with hip fractures, 0.60 days (95% C.I. 0.1–1.2) resulting in an observed percentage decrease of 8.3% (p value = 0.004). No other significant decrease in length of stay was observed in any of the other patient groups. The introduction of computer-aided triage and prioritization software into the radiological workflow was associated with a significant decrease in length of stay for patients diagnosed with ICH and PE.

Comparative Study of Diagnostic Efficacy of Single Phase-Computed Tomography Pulmonary Angiography and Dual Phase-Computed Tomography Pulmonary Angiography in the Diagnosis of Pulmonary Embolism

Objective

We compared the efficacy of single phase-computed tomography pulmonary angiography (SP-CTPA) and dual phase-computed tomography pulmonary angiography (DP-CTPA) for the diagnosis of pulmonary embolism (PE).

Methods

We recruited 1,019 consecutive patients (359 with PE) who underwent DP-CTPA (phase I: pulmonary artery phase; phase II: aortic phase) for suspected PE between January and October 2021. Phase I of DP-CTPA was used as SP-CTPA, and the final clinical diagnosis (FCD) was used as the gold standard.

Results

Three hundred fifty-two cases of PE were detected by both methods, with the same sensitivity of 98.1% (99.6–99.5%). Using SP-CTPA, 142 cases [13 pulmonary insufficiency artifacts (PIA) and 129 systemic-pulmonary shunt artifacts (S-PSA)] were false-positive with specificity of 78.5% (75.3–81.6%). No false-positive was found with DP-CTPA, with specificity of 100%, positive predictive value of 1, and negative predictive value of 0.990 (Net Reclassification Improvement = 0.215; P < 0.05). According to FCD, the positive results of SP-CTPA were divided into PIA, S-PSA, and true-positive (TP_SP−CTPA) groups, and pairwise comparisons were performed. The bronchiectasis and hemoptysis rate in S-PSA group was higher than that in PIA and TP groups (P < 0.001), and the pulmonary hypertension (PH) rate in PIA group was higher than that in S-PSA and TP groups (P < 0.001).

Conclusion

The diagnostic efficiency of DP-CTPA for the diagnosis of PE was high. SP-CTPA may misdiagnose PIA (common in patients with PH) and S-PSA (common in patients with bronchiectasis and hemoptysis) as PE.

Clinical importance of the distribution of pulmonary artery embolism in acute pulmonary embolism

Objective

To explore the clinical importance of the distribution of pulmonary artery embolism in acute pulmonary embolism (APE).

Methods

Sixty-four patients with APE were classified into mixed-type and distal-type pulmonary embolism groups. Their right ventricular systolic pressure (RVSP) and disease duration were recorded, and the diameter of their right ventricles was measured by ultrasound. The computed tomography angiographic clot load was determined as a Mastora score.

Results

Patients with distal-type pulmonary embolisms had significantly lower RVSPs (44.92 ± 17.04 vs 55.69 ± 17.66 mmHg), and significantly smaller right ventricular diameters (21.08 ± 3.06 vs 23.37 ± 3.48 mm) than those with mixed-type pulmonary embolisms. Additionally, disease duration was significantly longer in patients with distal-type pulmonary embolisms (14.33 ± 11.57 vs 8.10 ± 7.10 days), and they had significantly lower Mastora scores (20.91% ± 18.92% vs 43.96% ± 18.30%) than patients with mixed-type pulmonary embolisms. After treatment, RVSPs decreased significantly in patients with both distal-type and mixed-type pulmonary embolisms. Right ventricle diameters also decreased significantly in patients with mixed-type pulmonary embolisms after treatment.

Conclusion

Patients with mixed-type pulmonary embolisms are significantly more susceptible to pulmonary hypertension, enlarged right ventricular diameters, and shorter durations of disease than those with distal-type pulmonary embolisms. The distribution of pulmonary artery embolism in APE can provide a clinical reference.

Clinical Impact of a Radiologic Quality Initiative Promoting More Timely Communication of Critical Pulmonary Embolus Results

BACKGROUND

A section quality initiative was implemented beginning 2013 requiring positive pulmonary embolism (PE) results to be documented and communicated within 90 minutes of exam completion. The objective of this study is to evaluate the effect of this quality initiative on different intervals comprising the total patient processing time, namely the time from when the imaging exam was ordered to study completion interval, the time from study completion to positive PE result communication (TAT interval) or treatment initiation (TTT interval), the time from result communication to treatment initiation (TRCTI interval), and the total patient processing time (TPT interval).

METHODS

This was a retrospective, single-institution, IRB-approved cohort study that included 830 patients with the diagnosis of acute PE confirmed by CT pulmonary angiography. A maximum of 10 positive exams per month were identified and analyzed over an 84-month period from January 2010 to December 2016. The following data were obtained: time when exam ordered, time of imaging study completion, time of report completion, time of result communication, time of treatment, type of treatment, and reasons for any treatment delay. Analysis was done by determining the mean time spent in various intervals, the cumulative relative frequency of interval completion, and the fraction of the entire patient processing time spent in each interval.

RESULTS

Mean analysis demonstrated a decrease in all time intervals in the postpolicy period (ordered to study completion: Δ24.50%, p = 0.004; TAT: Δ23.91%, p < 0.001; TRCTI: Δ16.86%, p = 0.031; TTT: Δ17.40%, p = 0.005; TPT: Δ15.94%, p = 0.002). Cumulative relative frequency analysis demonstrated a higher rate of interval completion in the postpolicy period (TAT: p < 0.001; TRCTI: p = 0.007; TPT: p = 0.025). Interval fraction analysis demonstrated changes in the fraction of processing time spent in varying intervals (TAT: -Δ14.42%, p = 0.002; TRCTI: +Δ17.65%, p = 0.001).

CONCLUSION

Total patient processing time decreased after the policy implementation with a more significant decrease in TAT compared to other intervals. Radiologic processing time does not appear to be the rate-limiting step in total patient processing time.

Assessment of Acute Pulmonary Embolism by Computer-Aided Technique: A Reliability Study

Background

Acute pulmonary embolism is one of the most common cardiovascular diseases. Computer-aided technique is widely used in chest imaging, especially for assessing pulmonary embolism. The reliability and quantitative analyses of computer-aided technique are necessary. This study aimed to evaluate the reliability of geometry-based computer-aided detection and quantification for emboli morphology and severity of acute pulmonary embolism.

Material/Methods

Thirty patients suspected of acute pulmonary embolism were analyzed by both manual and computer-aided interpretation of vascular obstruction index and computer-aided measurements of emboli quantitative parameters. The reliability of Qanadli and Mastora scores was analyzed using computer-aided and manual interpretation.

Results

The time costs of manual and computer-aided interpretation were statistically different (374.90±150.16 versus 121.07±51.76, P<0.001). The difference between the computer-aided and manual interpretation of Qanadli score was 1.83±2.19, and 96.7% (29 out of 30) of the measurements were within 95% confidence interval (intraclass correlation coefficient, ICC=0.998). The difference between the computer-aided and manual interpretation of Mastora score was 1.46±1.62, and 96.7% (29 out of 30) of the measurements were within 95% confidence interval (ICC=0.997). The emboli quantitative parameters were moderately correlated with the Qanadli and Mastora scores (all P<0.001).

Conclusions

Computer-aided technique could reduce the time costs, improve the and reliability of vascular obstruction index and provided additional quantitative parameters for disease assessment.

Impact of an educational initiative targeting non-radiologist staff on overall notification times of critical findings in radiology

INTRODUCTION

The timely reporting of critical findings is considered by the Joint Commission as one of the main patient safety goals. Delays in critical radiological findings communication are directly related to delayed treatment initiation and death, constituting a major cause of medical malpractice suits. The aim of this study was to evaluate the impact of an educational initiative performed to reduce the notification times of critical radiological findings.

MATERIALS AND METHODS

All records of critical findings reported in the Radiology Department were evaluated. The notification times before and after performing the educational intervention taking into account the patient type, study, and critical diagnosis were calculated, evaluated, and compared. T test and chi-square test were used for statistical analysis, considering a p value less than 0.05 to indicate statistically significant differences.

RESULTS

We included 1949 reports, 805 before (41.3%) and 1144 (58.7%) after the intervention. Before the intervention, the mean time of critical finding report was 2.85 h for emergency patients and 3.07 h for hospitalized patients. After the intervention, a statistically significant decrease in the notification time was observed with a mean of 1.37 h for emergency patients and 2.43 h in the hospitalization patients. A statistically significant increase was observed in the proportion of reported findings in less than 15 min (7.08%, p < 0.01), 45 min (45.55%, p < 0.01), 60 min (55.86%, p < 0.01), and 120 min (80.68%, p < 0.01).

CONCLUSION

The healthcare process in the Department of Radiology involves multiple actors who must be sensitized in the identification and reporting of critical radiological findings in order to reduce the notification times. Ensuring effective communication of critical findings is indispensable to ensure timely medical treatment.

Global and Japanese regional variations in radiologist potential workload for computed tomography and magnetic resonance imaging examinations

PURPOSE

To investigate the global variation in radiologist potential workload for CT and MRI examinations, and the regional variation in potential workload and extent of radiologists' involvement in CT and MRI examinations in Japan.

METHODS

"Radiologist potential workload" was defined as the annual number of CT plus MRI examinations divided by the total number of diagnostic radiologists. The extent of radiologists' involvement was measured as the proportion of CT and MRI examinations to which "Added-fees for Radiological Managements on Imaging-studies (ARMIs)" were applied among eligible examinations. Maximum variation was computed as the ratio of the highest-to-lowest values among the countries or Japanese prefectures.

RESULTS

The radiologist potential workload in Japan was 2.78-4.17 times higher than those in other countries. A maximum prefecture-to-prefecture variation was 3.88. The average percentage of CT plus MRI examinations with ARMI applied was 43.3%, with a maximum prefecture-to-prefecture variation of 3.97. Prefectures with more radiologists tended to have a higher extent of radiologists' involvement.

CONCLUSIONS

Japan had a far greater radiologist potential workload compared with other countries, with a large regional variation among prefectures. Prefectures with more radiologists tended to have a higher extent of radiologists' involvement in CT and MRI examinations.

Computed Tomography Angiographic Assessment of Acute Chest Pain

Acute chest pain is a leading cause of Emergency Department visits. Computed tomography angiography plays a vital diagnostic role in such cases, but there are several common challenges associated with the imaging of acute chest pain, which, if unrecognized, can lead to an inconclusive or incorrect diagnosis. These imaging challenges fall broadly into 3 categories: (1) image acquisition, (2) image interpretation (including physiological and pathologic mimics), and (3) result communication. The aims of this review are to describe and illustrate the most common challenges in the imaging of acute chest pain and to provide solutions that will facilitate accurate diagnosis of the causes of acute chest pain in the emergency setting.

Radiologist involvement is associated with reduced use of MRI in the acute period of low back pain in a non-elderly population

PURPOSE

To test the hypothesis that "acute-period" lumbar MRI in non-elderly patients with low back pain is less frequently performed at clinics/hospitals with greater involvement of full-time radiologists in the imaging workflow.

METHODS

In a national-level claims database, we identified 14,819 non-elderly patients (mean age: 38.7±8.0 years) who visited clinics/hospitals for low back pain in 2013-2015. We classified the clinics/hospitals into four groups based on the level of full-time radiologist involvement and MRI ownership, and compared the frequency of acute-period lumbar MRI using hierarchical logistic regression analysis.

RESULTS

Patients visiting facilities without a full-time radiologist (n=2105) were significantly (p<0.001) more likely to undergo acute-period MRI than those visiting facilities with ≥1 radiologist partially managing imaging workflow (level-1, n=491) or ≥1 radiologist intensively involved in imaging workflow (level-2, n=1190) (15.7% vs. 6.9% and 7.3%; adjusted odds ratio of no-radiologist versus level-2: 2.93, p=0.018). No difference was observed between level-1 and level-2 involvement.

CONCLUSIONS

Facilities with no full-time radiologist were more likely to perform acute-period MRI to assess for low back pain, while no difference was seen between facilities with varying levels of radiologist involvement in the imaging workflow. Radiologist involvement may contribute to optimal utilisation of medical imaging.

KEY POINTS

• Lumbar MRI was more frequently performed at facilities without full-time radiologists. • Full-time radiologists may play an important role in appropriate utilisation of imaging. • Frequency of MRI was similar between moderate and intensive radiologist involvement.

Implementation and Performance of Automated Software for Computing Right-to-Left Ventricular Diameter Ratio From Computed Tomography Pulmonary Angiography Images

OBJECTIVE

The aim of this study was to prospectively test the performance and potential for clinical integration of software that automatically calculates the right-to-left ventricular (RV/LV) diameter ratio from computed tomography pulmonary angiography images.

METHODS

Using 115 computed tomography pulmonary angiography images that were positive for acute pulmonary embolism, we prospectively evaluated RV/LV ratio measurements that were obtained as follows: (1) completely manual measurement (reference standard), (2) completely automated measurement using the software, and (3 and 4) using a customized software interface that allowed 2 independent radiologists to manually adjust the automatically positioned calipers.

RESULTS

Automated measurements underestimated (P < 0.001) the reference standard (1.09 [0.25] vs1.03 [0.35]). With manual correction of the automatically positioned calipers, the mean ratio became closer to the reference standard (1.06 [0.29] by read 1 and 1.07 [0.30] by read 2), and the correlation improved (r = 0.675 to 0.872 and 0.887). The mean time required for manual adjustment (37 [20] seconds) was significantly less than the time required to perform measurements entirely manually (100 [23] seconds).

CONCLUSIONS

Automated CT RV/LV diameter ratio software shows promise for integration into the clinical workflow for patients with acute pulmonary embolism.

CT pulmonary angiography-based scoring system to predict the prognosis of acute pulmonary embolism

BACKGROUND

The purpose is to develop a comprehensive risk-scoring system based on CT findings for predicting 30-day mortality after acute pulmonary embolism (PE), and to compare it with PE Severity Index (PESI).

MATERIALS AND METHODS

The study included consecutive 1698 CT pulmonary angiograms (CTPA) positive for acute PE performed at a single institution (2003-2010). Two radiologists independently assessed each study regarding clinically relevant findings and then performed adjudication. These variables plus patient clinical information were included to build a LASSO logistic regression model to predict 30-day mortality. A point score for each significant variable was generated based on the final model. PESI score was calculated in 568 patients who visited the hospital after 2007.

RESULTS

Inter-reader agreements of interpretations were >95% except for septal bowing (92%). The final prediction model showed superior ability over PESI (AUC = 0.822 vs 0.745) for predicting all-cause 30-day mortality (12.4%). The scoring system based on the significant variables (age (years), pleural effusion (+20), pericardial effusion (+20), lung/liver/bone lesions suggesting malignancy (+60), chronic interstitial lung disease (+20), enlarged lymph node in thorax (+20), and ascites (+40)) stratified patients into 4 severity categories, with mortality rates of 0.008% in class-I (≤50 pt), 3.8% in class-II (51-100 pt), 17.6% in class-III (101-150 pt), and 40.9% in class-IV (>150 pt). The mortality rate in the CTPA-high risk category (class-IV) was higher than those in the PESI's high risk (27.4%) and very high risk (25.2%) categories.

CONCLUSION

The CTPA-based model was superior to PESI in predicting 30-day mortality. Incorporating the CTPA-based scoring system into image interpretation workflows may help physicians to select the most appropriate management approach for individual patients.

Association Between Confidence Level of Acute Pulmonary Embolism Diagnosis on CTPA images and Clinical Outcomes

RATIONALE AND OBJECTIVES

The purpose was to evaluate clinical characteristics associated with low confidence in diagnosis of acute pulmonary embolism (PE) as expressed in computed tomography pulmonary angiography (CTPA) reports and to evaluate the effect of confidence level in PE diagnosis on patient clinical outcomes.

MATERIALS AND METHODS

This study included radiology reports from 1664 consecutive CTPA considered positive for acute PE (8/2003-5/2010). All reports were retrospectively assessed for the level of confidence in diagnosis. Baseline characteristics and outcomes (therapies related to PE and short-term mortality) were compared between high and low confidence groups. Multivariable logistic and Cox regression analyses were used to analyze the relationship between the confidence level and outcomes.

RESULTS

One-hundred sixty of 1664 (9.6%) reports had language that reflected a low confidence in PE diagnosis. The low confidence group had smaller (segmental and subsegmental) suspected emboli (prevalence, 72.5% vs. 50.7%; P < .001) and more comorbidities. The low confidence group had a lower likelihood of receiving PE-related therapies (adjusted odds ratio [OR], 0.18; 95% confidence interval, 0.10-031, P < .001), but there was no change in the all-cause and PE-related 30-day and/or 90-day mortality (OR of death for low confidence, 0.81-1.13, P values > .5).

CONCLUSIONS

Roughly 10% of positive CTPA reports had uncertainty in PE findings, and patients with reports categorized as low confidence had smaller emboli and more comorbidities. Although the low confidence group was less likely to receive PE-related therapies, patients in this group were not associated with higher probability of short-term mortality.

Critical findings: timing of notification in neuroradiology

BACKGROUND AND PURPOSE

Timely reporting of critical findings in radiology has been identified by The Joint Commission as one of the National Patient Safety Goals. Our aim was to determine the magnitude of delays between identifying a neuroradiologic critical finding and verbally notifying the caregiver in an effort to improve clinical outcomes.

MATERIALS AND METHODS

We surveyed the time of critical finding discovery, attempted notification, and direct communication between neuroradiologists and caregivers for weekday, evening, overnight, and weekend shifts during an 8-week period. The data were collected by trained observers and/or trainees and included 13 neuroradiology attendings plus fellows and residents. Critical findings were based on a previously approved 17-item list. Summary and comparative t test statistics were calculated, and sources of delays were identified.

RESULTS

Ninety-one critical findings were recorded. The mean time from study acquisition to critical finding discovery was 62.2 minutes, from critical finding discovery to call made 3.7 minutes, and from call made to direct communication, 5.2 minutes. The overall time from critical finding discovery to caregiver notification was within 10 minutes in 72.5% (66/91) and 15 minutes in 93.4% (85/91) of cases. There were no significant differences across shifts except for daytime versus overnight and weekend shifts, when means were 2.4, 5.6, and 8.7 minutes, respectively (P < .01). If >1 physician was called, the mean notification time increased from 3.5 to 10.1 minutes (P < .01). Sources of delays included inaccurate contact information, physician unavailability (shift change/office closed), patient transfer to a different service, or lack of responsiveness from caregivers.

CONCLUSIONS

Direct communication with the responsible referring physician occurred consistently within 10-15 minutes after observation of a critical finding. These delays are less than the average interval from study acquisition to critical finding discovery (mean, 62.2 minutes).