RESUMO
OBJECTIVES: To assess whether a computer-aided detection (CADe) system could serve as a learning tool for radiology residents in chest X-ray (CXR) interpretation. METHODS: Eight radiology residents were asked to interpret 500 CXRs for the detection of five abnormalities, namely pneumothorax, pleural effusion, alveolar syndrome, lung nodule, and mediastinal mass. After interpreting 150 CXRs, the residents were divided into 2 groups of equivalent performance and experience. Subsequently, group 1 interpreted 200 CXRs from the "intervention dataset" using a CADe as a second reader, while group 2 served as a control by interpreting the same CXRs without the use of CADe. Finally, the 2 groups interpreted another 150 CXRs without the use of CADe. The sensitivity, specificity, and accuracy before, during, and after the intervention were compared. RESULTS: Before the intervention, the median individual sensitivity, specificity, and accuracy of the eight radiology residents were 43% (range: 35-57%), 90% (range: 82-96%), and 81% (range: 76-84%), respectively. With the use of CADe, residents from group 1 had a significantly higher overall sensitivity (53% [n = 431/816] vs 43% [n = 349/816], p < 0.001), specificity (94% [i = 3206/3428] vs 90% [n = 3127/3477], p < 0.001), and accuracy (86% [n = 3637/4244] vs 81% [n = 3476/4293], p < 0.001), compared to the control group. After the intervention, there were no significant differences between group 1 and group 2 regarding the overall sensitivity (44% [n = 309/696] vs 46% [n = 317/696], p = 0.666), specificity (90% [n = 2294/2541] vs 90% [n = 2285/2542], p = 0.642), or accuracy (80% [n = 2603/3237] vs 80% [n = 2602/3238], p = 0.955). CONCLUSIONS: Although it improves radiology residents' performances for interpreting CXRs, a CADe system alone did not appear to be an effective learning tool and should not replace teaching. CLINICAL RELEVANCE STATEMENT: Although the use of artificial intelligence improves radiology residents' performance in chest X-rays interpretation, artificial intelligence cannot be used alone as a learning tool and should not replace dedicated teaching. KEY POINTS: ⢠With CADe as a second reader, residents had a significantly higher sensitivity (53% vs 43%, p < 0.001), specificity (94% vs 90%, p < 0.001), and accuracy (86% vs 81%, p < 0.001), compared to residents without CADe. ⢠After removing access to the CADe system, residents' sensitivity (44% vs 46%, p = 0.666), specificity (90% vs 90%, p = 0.642), and accuracy (80% vs 80%, p = 0.955) returned to that of the level for the group without CADe.
Assuntos
Inteligência Artificial , Internato e Residência , Humanos , Raios X , Radiografia Torácica , RadiografiaRESUMO
OBJECTIVE: To compare computed tomography-enterography (CTE) and magnetic resonance-enterography (MRE) in the detection of right-sided bowel deep infiltrating endometriosis (DIE). MATERIALS AND METHODS: Fifty women with DIE who underwent preoperatively CTE and MRE were included. CTE and MRE were first analyzed separately by two independent readers who analyzed five bowel segments (cecum, appendix, ileocecal junction, distal ileum and proximal small bowel [i.e., proximal ileum and jejunum]) for the presence of DIE and then interpreted in consensus. CTE, MRE and CTE with MRE were compared in terms of sensitivity, specificity and accuracy. Interobserver agreement was assessed with kappa (κ) test. RESULTS: Using the reference standard 25 out 250 bowel segments were involved by DIE in 18 women and 225 were free of DIE. Sensitivity, specificity, and accuracy of CTE were 60% (95% confidence interval [CI]: 39-79), 93% (95% CI: 89-96) and 90% (95% CI: 85-93) for Reader 1, respectively, and 52% (95% CI: 31-72), 99% (95% CI: 97-100) and 94% (95% CI: 91-97) for Reader 2, with no differences in sensitivity (P = 0.564) and specificity (P = 0.181) between readers and fair interobserver agreement (κ = 0.37). For MRE these figures were 52% (95% CI: 31-72), 92% (95% CI: 88-95) and 88% (95% CI: 84-92) for Reader 1 and 60% (95% CI: 39-79), 99% (95% CI: 96-100) and 95% (95% CI: 91-97) for Reader 2, with no differences in sensitivity (P = 0.157) and specificity (P = 0.061) between readers and fair interobserver agreement (κ = 0.31). Significant differences in sensitivity (20%; 95% CI: 7-41) were found between CTE + MRE vs. CTE alone for Reader 1 and vs. MRE alone for Reader 2 (P = 0.041 for both) CONCLUSION: CTE and MRE have not different sensitivities and convey only fair interobserver agreement but are highly specific for the diagnosis of right-sided bowel DIE. CTE and MRE are complementary because they improve the detection of DIE implants when used in combination.
Assuntos
Endometriose , Humanos , Feminino , Endometriose/diagnóstico por imagem , Endometriose/cirurgia , Intestinos , Tomografia Computadorizada por Raios X , Intestino Delgado , Imageamento por Ressonância Magnética , Sensibilidade e EspecificidadeRESUMO
BACKGROUND: As the number of conventional radiographic examinations in pediatric emergency departments increases, so, too, does the number of reading errors by radiologists. OBJECTIVE: The aim of this study is to investigate the ability of artificial intelligence (AI) to improve the detection of fractures by radiologists in children and young adults. MATERIALS AND METHODS: A cohort of 300 anonymized radiographs performed for the detection of appendicular fractures in patients ages 2 to 21 years was collected retrospectively. The ground truth for each examination was established after an independent review by two radiologists with expertise in musculoskeletal imaging. Discrepancies were resolved by consensus with a third radiologist. Half of the 300 examinations showed at least 1 fracture. Radiographs were read by three senior pediatric radiologists and five radiology residents in the usual manner and then read again immediately after with the help of AI. RESULTS: The mean sensitivity for all groups was 73.3% (110/150) without AI; it increased significantly by almost 10% (P<0.001) to 82.8% (125/150) with AI. For junior radiologists, it increased by 10.3% (P<0.001) and for senior radiologists by 8.2% (P=0.08). On average, there was no significant change in specificity (from 89.6% to 90.3% [+0.7%, P=0.28]); for junior radiologists, specificity increased from 86.2% to 87.6% (+1.4%, P=0.42) and for senior radiologists, it decreased from 95.1% to 94.9% (-0.2%, P=0.23). The stand-alone sensitivity and specificity of the AI were, respectively, 91% and 90%. CONCLUSION: With the help of AI, sensitivity increased by an average of 10% without significantly decreasing specificity in fracture detection in a predominantly pediatric population.
Assuntos
Inteligência Artificial , Fraturas Ósseas , Adolescente , Adulto , Criança , Pré-Escolar , Fraturas Ósseas/diagnóstico por imagem , Humanos , Radiografia , Radiologistas , Estudos Retrospectivos , Adulto JovemRESUMO
BACKGROUND: Choroidal metastases are the most common eye metastatic site. The prevalence of choroidal metastases in NSCLC patients has been reported to vary from 0.2 to 7% in historical series. Although previously reported, little is known about choroidal metastasis in Epidermal Growth Factor Receptor (EGFR)-mutant Non-small cell lung cancer (NSCLC). This study sought to describe the prevalence of choroidal metastases among patients with EGFR-mutated NSCLC and their characteristics, and to estimate their impact on prognosis. METHODS: We conducted a single-center retrospective study including all consecutive metastatic EGFR-mutant NSCLC patients, from Sept. 2015 to Oct. 2018. The EGFR-mutant NSCLC patients were identified via the Department of Genetics' files. Patients who exhibited choroidal metastases were compared to patients without choroidal metastases. Kaplan-Meier analysis and log-rank test were conducted to assess median overall survival (OS) from diagnosis for the two groups. The study was approved by the IRB as CEPRO number #2020-010. RESULTS: Prevalence of choroidal metastases in EGFR-mutated NSCLCs was 8.4% (7/83). Five were women, and four current or former smokers. Molecular analysis showed three tumors with exon 19 deletion, three with L858R mutation, and one with complex exon 21 mutation. The choroidal metastases were symptomatic in six/seven patients. Visual disturbances decreased in all but one symptomatic cases upon EGFR TKI, and the choroidal response was maintained over time. Median follow-up was 42.2 mo (95%CI [37.2-47.1]). Median OS in the choroidal metastasis group was 23.4 mo (95%CI [0.1-51.4]) versus 27.9 mo (95%CI [16.9-38.9]) in the non-choroidal metastasis group (p = 0.32). In the choroidal metastasis group, 2-year and 5-year OS were 47.6 and 0%, respectively, versus 55.8 and 26.3% in the non-choroidal metastasis subset. CONCLUSIONS: Choroidal metastases in NSCLC EGFR-mutant patients are rare but should be systematically suspected in case of visual disturbance. TKIs are efficient for treating visual symptoms. Whether choroidal metastases confer a worse prognosis remains unclear owing to the third-generation EGFR TKI osimertinib first-line registration.