担当区分：最終著者,　責任著者  

Abstract

Background

The rapid advancement of generative artificial intelligence (AI) has revolutionized understanding and generation of human language. Their integration into healthcare has shown potential for improving medical diagnostics, yet a comprehensive diagnostic performance evaluation of generative AI models and the comparison of their diagnostic performance with that of physicians has not been extensively explored.

Methods

In this systematic review and meta-analysis, a comprehensive search of Medline, Scopus, Web of Science, Cochrane Central, and medRxiv was conducted for studies published from June 2018 through December 2023, focusing on those that validate generative AI models for diagnostic tasks. Meta-analysis was performed to summarize the performance of the models and to compare the accuracy of the models with that of physicians. The quality of studies was assessed using the Prediction Model Study Risk of Bias Assessment Tool.

Results

The search resulted in 54 studies being included in the meta-analysis, with 13 of these also used in the comparative analysis. Eight models were evaluated across 17 medical specialties. The overall accuracy for generative AI models across 54 studies was 57% (95% confidence interval [CI]: 51–63%). The I-squared statistic of 96% signifies a high degree of heterogeneity among the study results. Meta-regression analysis of generative AI models revealed significantly improved accuracy for GPT-4, and reduced accuracy for some specialties such as Neurology, Endocrinology, Rheumatology, and Radiology. The comparison meta-analysis demonstrated that, on average, physicians exceeded the accuracy of the models (difference in accuracy: 14% [95% CI: 8–19%], p-value <0.001). However, in the performance comparison between GPT-4 and physicians, GPT-4 performed slightly higher than non-experts (–4% [95% CI: –10–2%], p-value = 0.173), and slightly underperformed compared to experts (6% [95% CI: –1–13%], p-value = 0.091). The quality assessment indicated a high risk of bias in the majority of studies, primarily due to small sample sizes.

Conclusions

Generative AI exhibits promising diagnostic capabilities, with accuracy varying significantly by model and medical specialty. Although they have not reached the reliability of expert physicians, the findings suggest that generative AI models have the potential to enhance healthcare delivery and medical education, provided they are integrated with caution and their limitations are well-understood. This study also highlights the need for more rigorous research standards and a larger number of cases in the future.

DOI： 10.1101/2024.01.20.24301563

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）  

Abstract

Background

GPT-4-based ChatGPT demonstrates significant potential in various industries; however, its potential clinical applications remain largely unexplored.

Methods

We employed the New England Journal of Medicine (NEJM) quiz “Image Challenge” from October 2021 to March 2023 to assess ChatGPT's clinical capabilities. The quiz, designed for healthcare professionals, tests the ability to analyze clinical scenarios and make appropriate decisions. We evaluated ChatGPT's performance on the NEJM quiz, analyzing its accuracy rate by questioning type and specialty after excluding quizzes which were impossible to answer without images. ChatGPT was first asked to answer without the five multiple-choice options, and then after being given the options.

Results

ChatGPT achieved an 87% (54/62) accuracy without choices and a 97% (60/62) accuracy with choices, after excluding 16 image-based quizzes. Upon analyzing performance by quiz type, ChatGPT excelled in the Diagnosis category, attaining 89% (49/55) accuracy without choices and 98% (54/55) with choices. Although other categories featured fewer cases, ChatGPT's performance remained consistent. It demonstrated strong performance across the majority of medical specialties; however, Genetics had the lowest accuracy at 67% (2/3).

Conclusion

ChatGPT demonstrates potential for diagnostic applications, suggesting its usefulness in supporting healthcare professionals in making differential diagnoses and enhancing AI-driven healthcare.

DOI： 10.1186/s44247-023-00058-5

その他URL： https://link.springer.com/article/10.1186/s44247-023-00058-5/fulltext.html

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）  

DOI： 10.1016/s2589-7500(23)00224-8

PubMed

担当区分：最終著者,　責任著者  

Abstract

Objective

To compare the diagnostic accuracy of Generative Pre-trained Transformer (GPT)-4 based ChatGPT, GPT-4 with vision (GPT-4V) based ChatGPT, and radiologists in musculoskeletal radiology.

Materials and Methods

We included 106 “Test Yourself” cases fromSkeletal Radiologybetween January 2014 and September 2023. We input the medical history and imaging findings into GPT-4 based ChatGPT and the medical history and images into GPT-4V based ChatGPT, then both generated a diagnosis for each case. Two radiologists (a radiology resident and a board-certified radiologist) independently provided diagnoses for all cases. The diagnostic accuracy rates were determined based on the published ground truth. Chi-square tests were performed to compare the diagnostic accuracy of GPT-4 based ChatGPT, GPT-4V based ChatGPT, and radiologists.

Results

GPT-4 based ChatGPT significantly outperformed GPT-4V based ChatGPT (p< 0.001) with accuracy rates of 43% (46/106) and 8% (9/106), respectively. The radiology resident and the board-certified radiologist achieved accuracy rates of 41% (43/106) and 53% (56/106). The diagnostic accuracy of GPT-4 based ChatGPT was comparable to that of the radiology resident but was lower than that of the board-certified radiologist, although the differences were not significant (p= 0.78 and 0.22, respectively). The diagnostic accuracy of GPT-4V based ChatGPT was significantly lower than those of both radiologists (p< 0.001 and < 0.001, respectively).

Conclusion

GPT-4 based ChatGPT demonstrated significantly higher diagnostic accuracy than GPT-4V based ChatGPT. While GPT-4 based ChatGPT’s diagnostic performance was comparable to radiology residents, it did not reach the performance level of board-certified radiologists in musculoskeletal radiology.

DOI： 10.1101/2023.12.07.23299707

担当区分：最終著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

DOI： 10.1007/s00234-023-03252-4

PubMed

掲載種別：研究論文（学術雑誌）  

DOI： 10.1093/jrr/rrad090

PubMed

Abstract

Background

Although rotation atherectomy (RA) is a useful technique for severely calcified lesions, patients undergoing RA show a greater incidence of catastrophic complications, such as coronary perforation. Therefore, prior to the RA procedure, it is important to predict which regions of the coronary plaque will be debulked by RA.

Objectives

We develop and evaluate an artificial intelligence–based algorithm that uses pre-RA intravascular ultrasound (IVUS) images to automatically predict regions debulked by RA

Methods

A total of 2106 IVUS cross-sections from 60 patients with de novo severely calcified coronary lesions who underwent IVUS-guided RA were consecutively collected. The two identical IVUS images of pre-and post-RA were merged, and the orientations of the debulked segments identified in the merged images are marked on the outer circle of each IVUS image. The artificial intelligence model was developed based on ResNet (deep residual learning for image recognition). The architecture connected 36 fully connected layers, each corresponding to one of the 36 orientations segmented every 10°, to a single feature extractor.

Results

In each cross-sectional analysis, our artificial intelligence model achieved an average sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of 81%, 72%, 46%, 90%, and 75%, respectively.

Conclusions

The artificial intelligence–based algorithm can use information from pre-RA IVUS images to accurately predict regions debulked by RA. The proposed method will assist interventional cardiologists in determining the treatment strategies for severely calcified coronary lesions.

DOI： 10.1101/2023.11.07.23298239

担当区分：最終著者   掲載種別：研究論文（学術雑誌）  

DOI： 10.1016/j.jvir.2023.11.014

PubMed

担当区分：最終著者,　責任著者  

Abstract

Background

Large Language Models like Chat Generative Pre-trained Transformer (ChatGPT) have demonstrated potential for differential diagnosis in radiology. Previous studies investigating this potential primarily utilized quizzes from academic journals, which may not accurately represent real-world clinical scenarios.

Purpose

This study aimed to assess the diagnostic capabilities of ChatGPT using actual clinical radiology reports of brain tumors and compare its performance with that of neuroradiologists and general radiologists.

Methods

We consecutively collected brain MRI reports from preoperative brain tumor patients at Osaka Metropolitan University Hospital, taken from January to December 2021. ChatGPT and five radiologists were presented with the same findings from the reports and asked to suggest differential and final diagnoses. The pathological diagnosis of the excised tumor served as the ground truth. Chi-square tests and Fisher’s exact test were used for statistical analysis.

Results

In a study analyzing 99 radiological reports, ChatGPT achieved a final diagnostic accuracy of 75% (95% CI: 66, 83%), while radiologists’ accuracy ranged from 64% to 82%. ChatGPT’s final diagnostic accuracy using reports from neuroradiologists was higher at 82% (95% CI: 71, 89%), compared to 52% (95% CI: 33, 71%) using those from general radiologists with a p-value of 0.012. In the realm of differential diagnoses, ChatGPT’s accuracy was 95% (95% CI: 91, 99%), while radiologists’ fell between 74% and 88%. Notably, for these differential diagnoses, ChatGPT’s accuracy remained consistent whether reports were from neuroradiologists (96%, 95% CI: 89, 99%) or general radiologists (91%, 95% CI: 73, 98%) with a p-value of 0.33.

Conclusion

ChatGPT exhibited good diagnostic capability, comparable to neuroradiologists in differentiating brain tumors from MRI reports. ChatGPT can be a second opinion for neuroradiologists on final diagnoses and a guidance tool for general radiologists and residents, especially for understanding diagnostic cues and handling challenging cases.

Summary

This study evaluated ChatGPT’s diagnostic capabilities using real-world clinical MRI reports from brain tumor cases, revealing that its accuracy in interpreting brain tumors from MRI findings is competitive with radiologists.

Key results

ChatGPT demonstrated a diagnostic accuracy rate of 75% for final diagnoses based on preoperative MRI findings from 99 brain tumor cases, competing favorably with five radiologists whose accuracies ranged between 64% and 82%. For differential diagnoses, ChatGPT achieved a remarkable 95% accuracy, outperforming several of the radiologists.

Radiology reports from neuroradiologists and general radiologists showed varying accuracy when input into ChatGPT. Reports from neuroradiologists resulted in higher diagnostic accuracy for final diagnoses, while there was no difference in accuracy for differential diagnoses between neuroradiologists and general radiologists.

DOI： 10.1101/2023.10.27.23297585

掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

DOI： 10.1007/s11547-023-01691-w

PubMed

掲載種別：研究論文（学術雑誌）   国際・国内誌：国内誌  

DOI： 10.1007/s12149-023-01865-6

PubMed

担当区分：最終著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

DOI： 10.1016/s2666-7568(23)00133-2

PubMed

担当区分：最終著者,　責任著者  

Abstract

Purpose

To compare the diagnostic performance between Chat Generative Pre-trained Transformer (ChatGPT), based on the GPT-4 architecture, and radiologists from patient’s medical history and imaging findings in challenging neuroradiology cases.

Methods

We collected 30 consecutive “Freiburg Neuropathology Case Conference” cases from the journal Clinical Neuroradiology between March 2016 and June 2023. GPT-4 based ChatGPT generated diagnoses from the patient’s provided medical history and imaging findings for each case, and the diagnostic accuracy rate was determined based on the published ground truth. Three radiologists with different levels of experience (2, 4, and 7 years of experience, respectively) independently reviewed all the cases based on the patient’s provided medical history and imaging findings, and the diagnostic accuracy rates were evaluated. The Chi-square tests were performed to compare the diagnostic accuracy rates between ChatGPT and each radiologist.

Results

ChatGPT achieved an accuracy rate of 23% (7/30 cases). Radiologists achieved the following accuracy rates: a junior radiology resident had 27% (8/30) accuracy, a senior radiology resident had 30% (9/30) accuracy, and a board-certified radiologist had 47% (14/30) accuracy. ChatGPT’s diagnostic accuracy rate was lower than that of each radiologist, although the difference was not significant (p= 0.99, 0.77, and 0.10, respectively).

Conclusion

The diagnostic performance of GPT-4 based ChatGPT did not reach the performance level of either junior/senior radiology residents or board-certified radiologists in challenging neuroradiology cases. While ChatGPT holds great promise in the field of neuroradiology, radiologists should be aware of its current performance and limitations for optimal utilization.

DOI： 10.1101/2023.08.28.23294607

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国内誌  

DOI： 10.1007/s11604-023-01474-3

PubMed

担当区分：最終著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

DOI： 10.1148/radiol.223016

PubMed

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

DOI： 10.1016/s2589-7500(23)00107-3

PubMed

掲載種別：研究論文（学術雑誌）  

DOI： 10.1002/jmri.28892

PubMed

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

DOI： 10.1148/radiol.231040

PubMed

掲載種別：研究論文（学術雑誌）  

DOI： 10.1016/j.diii.2023.06.011

PubMed

担当区分：最終著者,　責任著者   掲載種別：研究論文（学術雑誌）  

DOI： 10.1183/16000617.0259-2022

PubMed

掲載種別：研究論文（学術雑誌）  

DOI： 10.1007/s11547-023-01638-1

PubMed

その他URL： https://link.springer.com/article/10.1007/s11547-023-01638-1/fulltext.html

担当区分：筆頭著者,　責任著者  

Background GPT-4-based ChatGPT demonstrates significant potential in various industries; however, its potential clinical applications remain largely unexplored. Methods We employed the New England Journal of Medicine (NEJM) quiz "Image Challenge" from October 2021 to March 2023 to assess ChatGPT's clinical capabilities. The quiz, designed for healthcare professionals, tests the ability to analyze clinical scenarios and make appropriate decisions. We evaluated ChatGPT's performance on the NEJM quiz, analyzing its accuracy rate by questioning type and specialty after excluding quizzes which were impossible to answer without images. The NEJM quiz has five multiple-choice options, but ChatGPT was first asked to answer without choices, and then given the choices to answer afterwards, in order to evaluate the accuracy in both scenarios. Results ChatGPT achieved an 87% accuracy without choices and a 97% accuracy with choices, after excluding 16 image-based quizzes. Upon analyzing performance by quiz type, ChatGPT excelled in the Diagnosis category, attaining 89% accuracy without choices and 98% with choices. Although other categories featured fewer cases, ChatGPT's performance remained consistent. It demonstrated strong performance across the majority of medical specialties; however, Genetics had the lowest accuracy at 67%. Conclusion ChatGPT demonstrates potential for clinical application, suggesting its usefulness in supporting healthcare professionals and enhancing AI-driven healthcare.

DOI： 10.1101/2023.05.04.23289493

掲載種別：研究論文（学術雑誌）  

DOI： 10.1016/j.jos.2023.05.004

PubMed

掲載種別：研究論文（学術雑誌）  

DOI： 10.3390/cancers15072140

PubMed

拡散テンソルイメージング(DTI)データの画像再構成により、analysis along the perivascular space(ALPS) indexにおいて再現性の改善を試みた。なお、本報では脳MRI画像データが登録されたオープンデータベースOASIS-3データセットから、認知機能が正常な234名(男性123名、女性111名、年齢42～92歳)の画像データを用いて実施した。また、DTIのベクトル情報を他の血管周囲腔に位置合わせする画像再構成技術を適用し、元データのALPS indexと再構成したALPS indexを新たに求め、再構成処理したdiffusivity mapを作成した。これらデータ間を比較した結果、再構成処理したALPS indexの変動は元データのALPS indexに比べ有意に小さく、再構成技術により、頭部回転を有する被験者でも良好ないし優れたALPS index数値の再現性が得られた。特にx軸で頭部回転が観察された被験者では、元データALPS indexと再構成したALPS indexを比較したBland-Altmanプロットにより、広範囲に及ぶ95%一致限界が認められたことから、x軸の頭部回転はALPS indexの算出に著しく影響を及ぼすことが示唆された。以上より、本法による再構成技術を適用したDTI-ALPS index算出によりALPSの再現性が改善され、再構成処理したdiffusivity mapも作成可能となった。

担当区分：最終著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

DOI： 10.1007/s10278-022-00691-y

PubMed

その他URL： https://link.springer.com/article/10.1007/s10278-022-00691-y/fulltext.html

DOI： 10.1148/radiol.232082

DOI： 10.11318/mii.40.66

掲載種別：研究論文（学術雑誌）   国際・国内誌：国内誌  

DOI： 10.2463/mrms.rev.2023-0047

担当区分：最終著者,　責任著者  

DOI： 10.48550/arXiv.2212.11113

掲載種別：研究論文（学術雑誌）  

DOI： 10.1007/s11604-022-01370-2

PubMed

その他URL： https://link.springer.com/article/10.1007/s11604-022-01370-2/fulltext.html

担当区分：最終著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

DOI： 10.1259/bjr.20220058

PubMed

担当区分：責任著者   掲載種別：研究論文（学術雑誌）  

<title>Abstract</title>We developed and validated a deep learning (DL)-based model using the segmentation method and assessed its ability to detect lung cancer on chest radiographs. Chest radiographs for use as a training dataset and a test dataset were collected separately from January 2006 to June 2018 at our hospital. The training dataset was used to train and validate the DL-based model with five-fold cross-validation. The model sensitivity and mean false positive indications per image (mFPI) were assessed with the independent test dataset. The training dataset included 629 radiographs with 652 nodules/masses and the test dataset included 151 radiographs with 159 nodules/masses. The DL-based model had a sensitivity of 0.73 with 0.13 mFPI in the test dataset. Sensitivity was lower in lung cancers that overlapped with blind spots such as pulmonary apices, pulmonary hila, chest wall, heart, and sub-diaphragmatic space (0.50–0.64) compared with those in non-overlapped locations (0.87). The dice coefficient for the 159 malignant lesions was on average 0.52. The DL-based model was able to detect lung cancers on chest radiographs, with low mFPI.

DOI： 10.1038/s41598-021-04667-w

その他URL： https://www.nature.com/articles/s41598-021-04667-w

症例は79歳女性。50歳代時に虫垂炎のため虫垂切除を行った。75歳時に子宮頸部扁平上皮細胞癌と診断され、放射線療法と化学療法を行い、治療の42ヵ月後の経過観察では再発は認められなかった。診察により左下部に圧痛が認められた。上腹部レントゲン検査でニボーを伴う弛緩性ループが観察され、当初は癒着性の小腸閉塞と診断された。保存療法では症状が変化しなかったため、腹部と骨盤の造影多列検出器コンピューター断層撮影を行った。その結果、S状結腸間膜の欠如による腸間膜脂肪組織と腸間膜血管による小腸への陥入が観察され、絞扼性小腸閉塞(TMSH)と診断した。緊急開腹手術を行い、TMSHが確認された。小腸の壊死部分を切除し、術後9日目に退院した。

担当区分：最終著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

OBJECTIVE: The purpose of this study was to develop an artificial intelligence (AI)-based model to detect features of atrial fibrillation (AF) on chest radiographs. METHODS: This retrospective study included consecutively collected chest radiographs of patients who had echocardiography at our institution from July 2016 to May 2019. Eligible radiographs had been acquired within 30 days of the echocardiography. These radiographs were labeled as AF-positive or AF-negative based on the associated electronic medical records; then, each patient was randomly divided into training, validation, and test datasets in an 8:1:1 ratio. A deep learning-based model to classify radiographs as with or without AF was trained on the training dataset, tuned with the validation dataset, and evaluated with the test dataset. RESULTS: The training dataset included 11,105 images (5637 patients; 3145 male, mean age ± standard deviation, 68 ± 14 years), the validation dataset included 1388 images (704 patients, 397 male, 67 ± 14 years), and the test dataset included 1375 images (706 patients, 395 male, 68 ± 15 years). Applying the model to the validation and test datasets gave a respective area under the curve of 0.81 (95% confidence interval, 0.78-0.85) and 0.80 (0.76-0.84), sensitivity of 0.76 (0.70-0.81) and 0.70 (0.64-0.76), specificity of 0.75 (0.72-0.77) and 0.74 (0.72-0.77), and accuracy of 0.75 (0.72-0.77) and 0.74 (0.71-0.76). CONCLUSION: Our AI can identify AF on chest radiographs, which provides a new way for radiologists to infer AF. KEY POINTS: • A deep learning-based model was trained to detect atrial fibrillation in chest radiographs, showing that there are indicators of atrial fibrillation visible even on static images. • The validation and test datasets each gave a solid performance with area under the curve, sensitivity, and specificity of 0.81, 0.76, and 0.75, respectively, for the validation dataset, and 0.80, 0.70, and 0.74, respectively, for the test dataset. • The saliency maps highlighted anatomical areas consistent with those reported for atrial fibrillation on chest radiographs, such as the atria.

DOI： 10.1007/s00330-022-08752-0

PubMed

その他URL： https://link.springer.com/article/10.1007/s00330-022-08752-0/fulltext.html

担当区分：責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

DOI： 10.1016/j.ejrad.2022.110433

PubMed

担当区分：責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

PURPOSE: To develop a deep learning (DL) model to generate synthetic, 2-dimensional subtraction angiograms free of artifacts from native abdominal angiograms. MATERIALS AND METHODS: In this retrospective study, 2-dimensional digital subtraction angiography (2D-DSA) images and native angiograms were consecutively collected from July 2019 to March 2020. Images were divided into motion-free (training, validation, and motion-free test datasets) and motion-artifact (motion-artifact test dataset) sets. A total of 3,185, 393, 383, and 345 images from 87 patients (mean age, 71 years ± 10; 64 men and 23 women) were included in the training, validation, motion-free, and motion-artifact test datasets, respectively. Native angiograms and 2D-DSA image pairs were used to train and validate an image-to-image translation model to generate synthetic DL-based subtraction angiography (DLSA) images. DLSA images were quantitatively evaluated by the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) using the motion-free dataset and were qualitatively evaluated via visual assessments by radiologists with a numerical rating scale using the motion-artifact dataset. RESULTS: The DLSA images showed a mean PSNR (± standard deviation) of 43.05 dB ± 3.65 and mean SSIM of 0.98 ± 0.01, indicating high agreement with the original 2D-DSA images in the motion-free dataset. Qualitative visual evaluation by radiologists of the motion-artifact dataset showed that DLSA images contained fewer motion artifacts than 2D-DSA images. Additionally, DLSA images scored similar to or higher than 2D-DSA images for vascular visualization and clinical usefulness. CONCLUSIONS: The developed DL model generated synthetic, motion-free subtraction images from abdominal angiograms with similar imaging characteristics to 2D-DSA images.

DOI： 10.1016/j.jvir.2022.03.010

PubMed

担当区分：責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国内誌  

OBJECTIVE: It is important to detect parathyroid adenomas by parathyroid scintigraphy with 99m-technetium sestamibi (99mTc-MIBI) before surgery. This study aimed to develop and validate deep learning (DL)-based models to detect parathyroid adenoma in patients with primary hyperparathyroidism, from parathyroid scintigrams with 99mTc-MIBI. METHODS: DL-based models for detecting parathyroid adenoma in early- and late-phase parathyroid scintigrams were, respectively, developed and evaluated. The training dataset used to train the models was collected from 192 patients (165 adenoma cases, mean age: 64 years ± 13, 145 women) and the validation dataset used to tune the models was collected from 45 patients (30 adenoma cases, mean age: 67 years ± 12, 37 women). The images were collected from patients who were pathologically diagnosed with parathyroid adenomas or in whom no lesions could be detected by either parathyroid scintigraphy or ultrasonography at our institution from June 2010 to March 2019. The models were tested on a dataset collected from 44 patients (30 adenoma cases, mean age: 67 years ± 12, 38 women) who took scintigraphy from April 2019 to March 2020. The models' lesion-based sensitivity and mean false positive indications per image (mFPI) were assessed with the test dataset. RESULTS: The sensitivity was 82% [95% confidence interval 72-92%] with mFPI of 0.44 for the scintigrams of the early-phase model and 83% [73-92%] with mFPI of 0.31 for the scintigrams of the delayed-phase model in the test dataset, respectively. CONCLUSIONS: The DL-based models were able to detect parathyroid adenomas with a high sensitivity using parathyroid scintigraphy with 99m-technetium sestamibi.

DOI： 10.1007/s12149-022-01726-8

PubMed

その他URL： https://link.springer.com/article/10.1007/s12149-022-01726-8/fulltext.html

副甲状腺機能亢進症患者の早期相ならびに後期相の99mTc-MIBIシンチグラフィー画像から副甲状腺腺腫を検出するため、深層学習(DL)をベースとしたモデルをそれぞれ開発し、検証した。DLモデルの学習データセットには患者192例(男性47例、女性145例、平均年齢64±13歳)から検出された165腺腫を選択し、検証データセットには、患者45例(男性8例、女性37例、平均年齢67±12歳)から検出された30腺腫とした。画像は2010年6月～2019年3月までの期間内に著者等の医療施設で、シンチグラフィーと超音波検査を受けた患者から収集し、試験データセットには、患者44例(男性6例、女性38例、平均年齢67±12歳)から検出された30腺腫を用いて解析した。その結果、初期相モデルによるシンチグラムでは病変をベースとした感度は82%[95%CI 72-92%]、1画像当りの平均偽陽性数(mFPI)は0.44で、後期相モデルではそれぞれ83%[95%CI 73-92%]、0.31であった。これらの所見から、99mTc-MIBIシンチグラフィー画像を用いて、本DLモデルを適用することにより、高感度に副甲状腺腺腫が検出されることが確認された。

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）  

Objectives

The objective of this study was to develop and validate a state-of-the-art, deep learning (DL)-based model for detecting breast cancers on mammography.

Methods

Mammograms in a hospital development dataset, a hospital test dataset, and a clinic test dataset were retrospectively collected from January 2006 through December 2017 in Osaka City University Hospital and Medcity21 Clinic. The hospital development dataset and a publicly available digital database for screening mammography (DDSM) dataset were used to train and to validate the RetinaNet, one type of DL-based model, with five-fold cross-validation. The model’s sensitivity and mean false positive indications per image (mFPI) and partial area under the curve (AUC) with 1.0 mFPI for both test datasets were externally assessed with the test datasets.

Results

The hospital development dataset, hospital test dataset, clinic test dataset, and DDSM development dataset included a total of 3179 images (1448 malignant images), 491 images (225 malignant images), 2821 images (37 malignant images), and 1457 malignant images, respectively. The proposed model detected all cancers with a 0.45–0.47 mFPI and had partial AUCs of 0.93 in both test datasets.

Conclusions

The DL-based model developed for this study was able to detect all breast cancers with a very low mFPI. Our DL-based model achieved the highest performance to date, which might lead to improved diagnosis for breast cancer.

DOI： 10.1371/journal.pone.0265751

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）  

DOI： 10.1148/ryai.210221

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）  

Abstract

Aims

We aimed to develop models to detect aortic stenosis (AS) from chest radiographs—one of the most basic imaging tests—with artificial intelligence.

Methods and results

We used 10 433 retrospectively collected digital chest radiographs from 5638 patients to train, validate, and test three deep learning models. Chest radiographs were collected from patients who had also undergone echocardiography at a single institution between July 2016 and May 2019. These were labelled from the corresponding echocardiography assessments as AS-positive or AS-negative. The radiographs were separated on a patient basis into training [8327 images from 4512 patients, mean age 65 ±  (standard deviation) 15 years], validation (1041 images from 563 patients, mean age 65 ± 14 years), and test (1065 images from 563 patients, mean age 65 ± 14 years) datasets. The soft voting-based ensemble of the three developed models had the best overall performance for predicting AS with an area under the receiver operating characteristic curve, sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of 0.83 (95% confidence interval 0.77–0.88), 0.78 (0.67–0.86), 0.71 (0.68–0.73), 0.71 (0.68–0.74), 0.18 (0.14–0.23), and 0.97 (0.96–0.98), respectively, in the validation dataset and 0.83 (0.78–0.88), 0.83 (0.74–0.90), 0.69 (0.66–0.72), 0.71 (0.68–0.73), 0.23 (0.19–0.28), and 0.97 (0.96–0.98), respectively, in the test dataset.

Conclusion

Deep learning models using chest radiographs have the potential to differentiate between radiographs of patients with and without AS.

Lay Summary

We created artificial intelligence (AI) models using deep learning to identify aortic stenosis (AS) from chest radiographs. Three AI models were developed and evaluated with 10 433 retrospectively collected radiographs and labelled from echocardiography reports. The ensemble AI model could detect AS in a test dataset with an area under the receiver operating characteristic curve of 0.83 (95% confidence interval 0.78–0.88). Since chest radiography is a cost-effective and widely available imaging test, our model can provide an additive resource for the detection of AS.

DOI： 10.1093/ehjdh/ztab102

その他URL： https://academic.oup.com/ehjdh/article-pdf/3/1/20/47116703/ztab102.pdf

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

<sec><title>PURPOSE</title> The molecular subtype of breast cancer is an important component of establishing the appropriate treatment strategy. In clinical practice, molecular subtypes are determined by receptor expressions. In this study, we developed a model using deep learning to determine receptor expressions from mammograms.

</sec><sec><title>METHODS</title> A developing data set and a test data set were generated from mammograms from the affected side of patients who were pathologically diagnosed with breast cancer from January 2006 through December 2016 and from January 2017 through December 2017, respectively. The developing data sets were used to train and validate the DL-based model with five-fold cross-validation for classifying expression of estrogen receptor (ER), progesterone receptor (PgR), and human epidermal growth factor receptor 2-neu (HER2). The area under the curves (AUCs) for each receptor were evaluated with the independent test data set.

</sec><sec><title>RESULTS</title> The developing data set and the test data set included 1,448 images (997 ER-positive and 386 ER-negative, 641 PgR-positive and 695 PgR-negative, and 220 HER2-enriched and 1,109 non–HER2-enriched) and 225 images (176 ER-positive and 40 ER-negative, 101 PgR-positive and 117 PgR-negative, and 53 HER2-enriched and 165 non–HER2-enriched), respectively. The AUC of ER-positive or -negative in the test data set was 0.67 (0.58-0.76), the AUC of PgR-positive or -negative was 0.61 (0.53-0.68), and the AUC of HER2-enriched or non–HER2-enriched was 0.75 (0.68-0.82).

</sec><sec><title>CONCLUSION</title> The DL-based model effectively classified the receptor expressions from the mammograms. Applying the DL-based model to predict breast cancer classification with a noninvasive approach would have additive value to patients.

</sec>

DOI： 10.1200/PO.20.00176

PubMed

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

<title>Abstract</title><sec>
<title>Background</title>
We investigated the performance improvement of physicians with varying levels of chest radiology experience when using a commercially available artificial intelligence (AI)-based computer-assisted detection (CAD) software to detect lung cancer nodules on chest radiographs from multiple vendors.


</sec><sec>
<title>Methods</title>
Chest radiographs and their corresponding chest CT were retrospectively collected from one institution between July 2017 and June 2018. Two author radiologists annotated pathologically proven lung cancer nodules on the chest radiographs while referencing CT. Eighteen readers (nine general physicians and nine radiologists) from nine institutions interpreted the chest radiographs. The readers interpreted the radiographs alone and then reinterpreted them referencing the CAD output. Suspected nodules were enclosed with a bounding box. These bounding boxes were judged correct if there was significant overlap with the ground truth, specifically, if the intersection over union was 0.3 or higher. The sensitivity, specificity, accuracy, PPV, and NPV of the readers’ assessments were calculated.


</sec><sec>
<title>Results</title>
In total, 312 chest radiographs were collected as a test dataset, including 59 malignant images (59 nodules of lung cancer) and 253 normal images. The model provided a modest boost to the reader’s sensitivity, particularly helping general physicians. The performance of general physicians was improved from 0.47 to 0.60 for sensitivity, from 0.96 to 0.97 for specificity, from 0.87 to 0.90 for accuracy, from 0.75 to 0.82 for PPV, and from 0.89 to 0.91 for NPV while the performance of radiologists was improved from 0.51 to 0.60 for sensitivity, from 0.96 to 0.96 for specificity, from 0.87 to 0.90 for accuracy, from 0.76 to 0.80 for PPV, and from 0.89 to 0.91 for NPV. The overall increase in the ratios of sensitivity, specificity, accuracy, PPV, and NPV were 1.22 (1.14–1.30), 1.00 (1.00–1.01), 1.03 (1.02–1.04), 1.07 (1.03–1.11), and 1.02 (1.01–1.03) by using the CAD, respectively.


</sec><sec>
<title>Conclusion</title>
The AI-based CAD was able to improve the ability of physicians to detect nodules of lung cancer in chest radiographs. The use of a CAD model can indicate regions physicians may have overlooked during their initial assessment.


</sec>

DOI： 10.1186/s12885-021-08847-9

PubMed

その他URL： https://link.springer.com/article/10.1186/s12885-021-08847-9/fulltext.html

掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

BACKGROUND AND AIMS: We developed a deep learning (DL) model for automated atherosclerotic plaque categorization using optical frequency domain imaging (OFDI) and performed quantitative and visual evaluations. METHODS: A total of 1103 histological cross-sections from 45 autopsy hearts were examined to compare the ex vivo OFDI scans. The images were segmented and annotated considering four histological categories: pathological intimal thickening (PIT), fibrous cap atheroma (FA), fibrocalcific plaque (FC), and healed erosion/rupture (HER). The DL model was developed based on pyramid scene parsing network (PSPNet). Given an input image, a convolutional neural network (ResNet50) was used as an encoder to generate feature maps of the last convolutional layer. RESULTS: For the quantitative evaluation, the mean F-score and IoU values, which are used to evaluate how close the predicted results are to the ground truth, were used. The validation and test dataset had F-score and IoU values of 0.63, 0.49, and 0.66, 0.52, respectively. For the section-level diagnostic accuracy, the areas under the receiver-operating characteristic curve produced by the DL model for FC, PIT, FA, and HER were 0.91, 0.85, 0.86, and 0.86, respectively, and were comparable to those of an expert observer. CONCLUSIONS: DL semantic segmentation of coronary plaques in OFDI images was used as a tool to automatically categorize atherosclerotic plaques using histological findings as the gold standard. The proposed method can support interventional cardiologists in understanding histological properties of plaques.

DOI： 10.1016/j.atherosclerosis.2021.06.003

PubMed

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

Background Digital subtraction angiography (DSA) generates an image by subtracting a mask image from a dynamic angiogram. However, patient movement-caused misregistration artifacts can result in unclear DSA images that interrupt procedures. Purpose To train and to validate a deep learning (DL)-based model to produce DSA-like cerebral angiograms directly from dynamic angiograms and then quantitatively and visually evaluate these angiograms for clinical usefulness. Materials and Methods A retrospective model development and validation study was conducted on dynamic and DSA image pairs consecutively collected from January 2019 through April 2019. Angiograms showing misregistration were first separated per patient by two radiologists and sorted into the misregistration test data set. Nonmisregistration angiograms were divided into development and external test data sets at a ratio of 8:1 per patient. The development data set was divided into training and validation data sets at ratio of 3:1 per patient. The DL model was created by using the training data set, tuned with the validation data set, and then evaluated quantitatively with the external test data set and visually with the misregistration test data set. Quantitative evaluations used the peak signal-to-noise ratio (PSNR) and the structural similarity (SSIM) with mixed liner models. Visual evaluation was conducted by using a numerical rating scale. Results The training, validation, nonmisregistration test, and misregistration test data sets included 10 751, 2784, 1346, and 711 paired images collected from 40 patients (mean age, 62 years ± 11 [standard deviation]; 33 women). In the quantitative evaluation, DL-generated angiograms showed a mean PSNR value of 40.2 dB ± 4.05 and a mean SSIM value of 0.97 ± 0.02, indicating high coincidence with the paired DSA images. In the visual evaluation, the median ratings of the DL-generated angiograms were similar to or better than those of the original DSA images for all 24 sequences. Conclusion The deep learning-based model provided clinically useful cerebral angiograms free from clinically significant artifacts directly from dynamic angiograms. Published under a CC BY 4.0 license. Supplemental material is available for this article.

DOI： 10.1148/radiol.2021203692

PubMed

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国内誌  

PURPOSE: To demonstrate how artificial intelligence (AI) can expand radiologists' capacity, we visualized the features of invasive ductal carcinomas (IDCs) that our algorithm, developed and validated for basic pathological classification on mammograms, had focused on. MATERIALS AND METHODS: IDC datasets were built using mammograms from patients diagnosed with IDCs from January 2006 to December 2017. The developing dataset was used to train and validate a VGG-16 deep learning (DL) network. The true positives (TPs) and accuracy of the algorithm were externally evaluated using the test dataset. A visualization technique was applied to the algorithm to determine which malignant findings on mammograms were revealed. RESULTS: The datasets were split into a developing dataset (988 images) and a test dataset (131 images). The proposed algorithm diagnosed 62 TPs with an accuracy of 0.61-0.70. The visualization of features on the mammograms revealed that the tubule forming, solid, and scirrhous types of IDCs exhibited visible features on the surroundings, corners of the masses, and architectural distortions, respectively. CONCLUSION: We successfully showed that features isolated by a DL-based algorithm trained to classify IDCs were indeed those known to be associated with each pathology. Thus, using AI can expand the capacity of radiologists through the discovery of previously unknown findings.

DOI： 10.1007/s11604-020-01070-9

PubMed

その他URL： http://link.springer.com/article/10.1007/s11604-020-01070-9/fulltext.html

マンモグラム上での浸潤性乳管癌(IDC)の病理分類のため、深層学習(DL)をベースとしたアルゴリズムの開発と検証に人工知能(AI)を適用し、その有用性について検討した。なお、学習データセットは2006年1月～2016年12月までの期間内に、当院で乳癌と診断された患者529例(年齢25～97歳)から得たマンモグラフィ画像988画像を用いて作成した。VGG-16DLネットワークにより、学習データセットと検証データセットを作成し、テストデータセットには2017年1月～2017年12月までの期間内に、乳癌と診断された当院患者67例(年齢40～92歳)から得た131画像を用いて作成し、アルゴリズムの真陽性と正確さを判定した。可視化技術をアルゴリズムに適用し、マンモグラム上で判明した悪性所見を調査した結果、DLベースのアルゴリズムにより、62画像の真陽性が0.61～0.70の正確さで診断され、マンモグラム上で認められた特徴の可視化により、腺管形成型、充実型および硬性型のIDCが、それぞれ腫瘤の周囲組織や乳腺の歪み、構築の乱れ等、可視化される特徴で描出されることが確認された。以上の所見から、DLベースのアルゴリズムにより、各分類でIDC病理との関連性が指摘されている画像特徴が抽出され、AIの適用により、これまで不明であった所見が明らかにされた。

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国内誌  

Deep learning has been applied to clinical applications in not only radiology, but also all other areas of medicine. This review provides a technical and clinical overview of deep learning in radiology. To gain a more practical understanding of deep learning, deep learning techniques are divided into five categories: classification, object detection, semantic segmentation, image processing, and natural language processing. After a brief overview of technical network evolutions, clinical applications based on deep learning are introduced. The clinical applications are then summarized to reveal the features of deep learning, which are highly dependent on training and test datasets. The core technology in deep learning is developed by image classification tasks. In the medical field, radiologists are specialists in such tasks. Using clinical applications based on deep learning would, therefore, be expected to contribute to substantial improvements in radiology. By gaining a better understanding of the features of deep learning, radiologists could be expected to lead medical development.

DOI： 10.1007/s11604-018-0795-3

PubMed

その他URL： http://link.springer.com/content/pdf/10.1007/s11604-018-0795-3.pdf

担当区分：筆頭著者,　責任著者   掲載種別：研究論文（学術雑誌）   国際・国内誌：国際誌  

Purpose To develop and evaluate a supportive algorithm using deep learning for detecting cerebral aneurysms at time-of-flight MR angiography to provide a second assessment of images already interpreted by radiologists. Materials and Methods MR images reported by radiologists to contain aneurysms were extracted from four institutions for the period from November 2006 through October 2017. The images were divided into three data sets: training data set, internal test data set, and external test data set. The algorithm was constructed by deep learning with the training data set, and its sensitivity to detect aneurysms in the test data sets was evaluated. To find aneurysms that had been overlooked in the initial reports, two radiologists independently performed a blinded interpretation of aneurysm candidates detected by the algorithm. When there was disagreement, the final diagnosis was made in consensus. The number of newly detected aneurysms was also evaluated. Results The training data set, which provided training and validation data, included 748 aneurysms (mean size, 3.1 mm ± 2.0 [standard deviation]) from 683 examinations; 318 of these examinations were on male patients (mean age, 63 years ± 13) and 365 were on female patients (mean age, 64 years ± 13). Test data were provided by the internal test data set (649 aneurysms [mean size, 4.1 mm ± 3.2] in 521 examinations, including 177 male patients and 344 female patients with mean age of 66 years ± 12 and 67 years ± 13, respectively) and the external test data set (80 aneurysms [mean size, 4.1 mm ± 2.1] in 67 examinations, including 19 male patients and 48 female patients with mean age of 63 years ± 12 and 68 years ± 12, respectively). The sensitivity was 91% (592 of 649) and 93% (74 of 80) for the internal and external test data sets, respectively. The algorithm improved aneurysm detection in the internal and external test data sets by 4.8% (31 of 649) and 13% (10 of 80), respectively, compared with the initial reports. Conclusion A deep learning algorithm detected cerebral aneurysms in radiologic reports with high sensitivity and improved aneurysm detection compared with the initial reports. © RSNA, 2018 See also the editorial by Flanders in this issue.

DOI： 10.1148/radiol.2018180901

PubMed

2019年9月17日、医薬品医療機器総合機構(PMDA)は、大阪市立大学大学院医学研究科放射線診断学・IVR学教室(以下、当教室)が共同開発したmagnetic resonance angiography(MRA)からの脳動脈瘤補助診断プログラム(EIRL Aneurysm)を、人工知能(artificial intelligence:AI)の中でも特にディープラーニングを用いた医療機器として、日本で初めて医療機器として承認した。これは、まさに医療分野におけるAI時代の幕開けといえる。本稿では、「AI時代の新医療」について、AIと人間の認知の差異の観点を軸にしつつ、当教室でのAI研究の実例を交えて考えていきたい。(著者抄録)

近年、AIの放射線分野への応用が進んでいる。画像診断支援分野におけるAI分野の研究を大別すると、分類、検出、セグメンテーション、画像処理の4つとなる。これらそれぞれの分野を筆者らの研究を交えつつ紹介した。さらに、ディープラーニングで最近注目されている技術、および、AIと医療者の認知の違いについての実験を紹介し、来たるAI時代の医療者の役割について考察した。

J-GLOBAL

2019年9月17日、医薬品医療機器総合機構(PMDA)は、大阪市立大学大学院医学研究科放射線診断学・IVR学教室(以下、当教室)が共同開発したmagnetic resonance angiography(MRA)からの脳動脈瘤補助診断プログラム(EIRL Aneurysm)を、人工知能(artificial intelligence:AI)の中でも特にディープラーニングを用いた医療機器として、日本で初めて医療機器として承認した。これは、まさに医療分野におけるAI時代の幕開けといえる。本稿では、「AI時代の新医療」について、AIと人間の認知の差異の観点を軸にしつつ、当教室でのAI研究の実例を交えて考えていきたい。(著者抄録)

J-GLOBAL

J-GLOBAL

掲載種別：研究発表ペーパー・要旨（全国大会，その他学術会議）  

J-GLOBAL

J-GLOBAL

会議種別：ポスター発表  

会議種別：口頭発表（招待・特別）  

会議種別：口頭発表（招待・特別）  

会議種別：口頭発表（招待・特別）  

会議種別：口頭発表（招待・特別）  

会議種別：口頭発表（招待・特別）  

会議種別：口頭発表（基調）  

会議種別：公開講演，セミナー，チュートリアル，講習，講義等  

会議種別：公開講演，セミナー，チュートリアル，講習，講義等  

会議種別：公開講演，セミナー，チュートリアル，講習，講義等