Initially Missed or Delayed Diagnosis of Subarachnoid Hemorrhage: A Nationwide Survey of Contributing Factors and Outcomes in Japan

Artificial Intelligence Trained by Deep Learning Can Improve Computed Tomography Diagnosis of Nontraumatic Subarachnoid Hemorrhage by Nonspecialists

Subarachnoid hemorrhage (SAH) is a serious cerebrovascular disease with a high mortality rate and is known as a disease that is hard to diagnose because it may be overlooked by noncontrast computed tomography (NCCT) examinations that are most frequently used for diagnosis. To create a system preventing this oversight of SAH, we trained artificial intelligence (AI) with NCCT images obtained from 419 patients with nontraumatic SAH and 338 healthy subjects and created an AI system capable of diagnosing the presence and location of SAH. Then, we conducted experiments in which five neurosurgery specialists, five nonspecialists, and the AI system interpreted NCCT images obtained from 135 patients with SAH and 196 normal subjects. The AI system was capable of performing a diagnosis of SAH with equal accuracy to that of five neurosurgery specialists, and the accuracy was higher than that of nonspecialists. Furthermore, the diagnostic accuracy of four out of five nonspecialists improved by interpreting NCCT images using the diagnostic results of the AI system as a reference, and the number of oversight cases was significantly reduced by the support of the AI system. This is the first report demonstrating that an AI system improved the diagnostic accuracy of SAH by nonspecialists.

Ruptured intracranial aneurysm presenting as cerebral circulation insufficiency: A case report

BACKGROUND

Rupture of an intracranial aneurysm is a type of acute stroke that is a serious threat to human health. Misdiagnosis of ruptured intracranial aneurysms is a serious clinical event that may have disastrous consequences in some patients. To date, ruptured intracranial aneurysms have been misdiagnosed as meningitis, tumors, stroke, or trauma, among other conditions. Here, we report what appears to be the first case of a ruptured intracranial aneurysm that presented as cerebral circulation insufficiency.

CASE SUMMARY

A middle-aged man was admitted to our hospital because of a parasellar lesion identified on a noncontrast computed tomography (CT) image after a mild traffic accident that was caused by a brief loss of consciousness. Notably, he was diagnosed with cerebral circulation insufficiency after two unexplained episodes of a transient loss of consciousness within the past 8 mo. The patient was diagnosed with right internal carotid artery aneurysm based on CT angiography and completely recovered after a craniotomy at our hospital. A few clots and severe adhesions around the aneurysm were observed in the subarachnoid space during the operation, suggesting that the aneurysm had ruptured and may had been misdiagnosed as cerebral circulation insufficiency.

CONCLUSION

Ruptured intracranial aneurysms may show negative imaging results and present as cerebral circulation insufficiency, which should be recognized as soon as possible to ensure timely management.

Improvement of the diagnostic accuracy for intracranial haemorrhage using deep learning-based computer-assisted detection

PURPOSE

To elucidate the effect of deep learning-based computer-assisted detection (CAD) on the performance of different-level physicians in detecting intracranial haemorrhage using CT.

METHODS

A total of 40 head CT datasets (normal, 16; haemorrhagic, 24) were evaluated by 15 physicians (5 board-certificated radiologists, 5 radiology residents, and 5 medical interns). The physicians attended 2 reading sessions without and with CAD. All physicians annotated the haemorrhagic regions with a degree of confidence, and the reading time was recorded in each case. Our CAD system was developed using 433 patients' head CT images (normal, 203; haemorrhagic, 230), and haemorrhage rates were displayed as corresponding probability heat maps using U-Net and a machine learning-based false-positive removal method. Sensitivity, specificity, accuracy, and figure of merit (FOM) were calculated based on the annotations and confidence levels.

RESULTS

In patient-based evaluation, the mean accuracy of all physicians significantly increased from 83.7 to 89.7% (p < 0.001) after using CAD. Additionally, accuracies of board-certificated radiologists, radiology residents, and interns were 92.5, 82.5, and 76.0% without CAD and 97.5, 90.5, and 81.0% with CAD, respectively. The mean FOM of all physicians increased from 0.78 to 0.82 (p = 0.004) after using CAD. The reading time was significantly lower when CAD (43 s) was used than when it was not (68 s, p < 0.001) for all physicians.

CONCLUSION

The CAD system developed using deep learning significantly improved the diagnostic performance and reduced the reading time among all physicians in detecting intracranial haemorrhage.

Evaluation of subacute subarachnoid haemorrhage detection using a magnetic resonance imaging sequence: Double inversion recovery

BACKGROUND AND OBJECTIVES

The diagnosis of subarachnoid hemorrhage (SAH) especially at the subacute stage is still a challenging issue using the conventional imaging modalities. Here we evaluated the role of double inversion recovery (DIR) sequence of MRI compared with the conventional gradient-recalled echo (GRE)-T2*-W and susceptibility-weighted imaging (SWI) sequences in the diagnosis of subacute SAH.

MATERIALS AND METHODS

This prospective study was conducted on 21 patients with SAH, which were diagnosed using CT scan at the initial step. In the third week after the injury (14-20 days), all patients underwent a brain MRI exam that included T2*-W, SWI, and DIR imaging sequences. All images were independently read by two radiologists, who were blinded to the clinical history of the patients. The presence or absence of SAH was reviewed and assessed in 6 anatomical regions.

RESULTS

On the DIR images, 20 patients were found to have at least one subarachnoid signal abnormality, while the SWI and T2*-W images identified SAH areas on 17 and 15 patients, respectively. The highest rate of inter-observer consensus by the DIR sequence was found in the interhemispheric fissure and perimesencephalic area (k = 1). Also, a highest rate of inter-observer consensus using SWI was found in the interhemispheric fissure and posterior fossa cistern area (k = 1). A weak agreement was found in frontal-parietal convexity using SWI (k = 0.447), and in posterior fossa cistern by the T2* sequence (k = 0.447).

CONCLUSION

In conclusion, the DIR sequence was more reliable at identifying signal abnormalities in subacute SAH patients than the T2*-W and SWI sequence, and is suggested as a promising imaging technique for detecting hemorrhagic areas without considering the anatomical distribution of SAH.

Subarachnoid Hemorrhage and Headache

PURPOSE FOR REVIEW

Subarachnoid hemorrhage is a serious and life-threatening medical condition which commonly presents with an acute headache. Unfortunately, it remains frequently misdiagnosed at initial presentation with dire consequences in terms of patient morbidity and mortality. The goal of this paper is to review salient features in the clinical history, as well as recently developed clinical decision rules, which can help determine which patients warrant further investigation for subarachnoid hemorrhage when the initial presentation is that of an acute headache.

RECENT FINDINGS

A recent prospective observational study showed that occipital location, stabbing quality, presence of meningism, and onset of headache during exertion were characteristics in the clinical history that can distinguish the headache of SAH from other causes. The Ottawa headache rule is a clinical decision tool which was developed to help identify patients presenting to the ED with acute non-traumatic headache who require investigation to rule out subarachnoid hemorrhage. Using this tool, it is recommended that patients who meet any one of the following 6 criteria are investigated further: Onset greater than or equal to 40 years, presence of neck pain or stiffness, witnessed loss of consciousness, onset during exertion, thunder clap headache (pain peaking within 1 s), or limited neck flexion on exam. An informed and thoughtful approach that takes into account the timing, presentation, risk factors, and resources, as discussed here, should help distinguish between the patient that warrants further evaluation and intervention for SAH and one who does not. The Ottawa SAH rule is a useful clinical decision tool for young inexperienced clinicians in order to avoid missed diagnoses. However, its clinical value is limited by its low specificity. Clinical decision tools with higher specificity are needed.