Deep Learning-Based Functional Independence Measure Score Prediction After Stroke in Kaifukuki (Convalescent) Rehabilitation Ward Annexed to Acute Care Hospital.

Introduction Prediction models of functional independent measure (FIM) score after kaifukuki (convalescent) rehabilitation ward (KRW) are needed to decide the treatment strategies and save medical resources. Statistical models were reported, but their accuracies were not satisfactory. We made such prediction models using the deep learning (DL) framework, Prediction One (Sony Network Communications Inc., Tokyo, Japan). Methods Of the 559 consecutive stroke patients, 122 patients were transferred to our KRW. We divided our 122 patients' data randomly into halves of training and validation datasets. Prediction One made three prediction models from the training dataset using (1) variables at the acute care ward admission, (2) those at the KRW admission, and (3) those combined (1) and (2). The models' determination coefficients (R2), correlation coefficients (rs), and residuals were calculated using the validation dataset. Results Of the 122 patients, the median age was 71, length of stay (LOS) in acute care ward 23 (17-30) days, LOS in KRW 53 days, total FIM scores at the admission of KRW 85, those at discharge 108. The mean FIM gain and FIM efficiency were 19 and 0.417. All patients were discharged home. Model (1), (2), and (3)'s R2 were 0.794, 0.970, and 0.972. Their mean residuals between the predicted and actual total FIM scores were -1.56±24.6, -4.49±17.1, and -2.69±15.7. Conclusion Our FIM gain and efficiency were better than national averages of FIM gain 17.1 and FIM efficiency 0.187. We made DL-based total FIM score prediction models, and their accuracies were superior to those of previous statistically calculated ones. The DL-based FIM score prediction models would save medical costs and perform efficient stroke and rehabilitation medicine.

Chronic subdural hematoma in patients over 65 years old: Results of using a postoperative cognitive evaluation to determine whether to permit return to driving.

BACKGROUND: Chronic subdural hematoma (CSDH) is usually associated with good recovery with burr hole irrigation and postoperative drainage under local anesthesia. In Japan, traffic accidents by the elderly drivers over 65 years old are severely increasing, and there is no consensus on whether or not to return to driving after CSDH treatment. We perform a postoperative cognitive assessment. We retrospectively investigated the return-to-driving rate and associated factors. METHODS: Of the 45 patients over 65 y.o. and who had usually driven, 30 patients wished to drive again. We performed tests composed of Mini-Mental State Examination (MMSE), line cancellation and line bisection task, Kohs block design test, trail making test (TMT)-A and B, Kana-hiroi test, Rey-Osterrieth complex figure test, and behavioral assessment of the dysexecutive syndrome, in order. When all tests' scores were better than the cutoff values, we let patients drive again. When some of the scores were worse than the cutoff values, we reevaluated the patients at the outpatient every month. If the patients' scores could not improve at the outpatient, we recommended them to stop driving. RESULTS: Nineteen of 30 patients could return to driving. Worse MMSE, Kohs block design test, TMT-A, TMT-B scores, higher age, dementia, or consciousness disturbance as chief complaints were associated with driving disability. CONCLUSION: CSDH is known as treatable dementia. However, we should perform an objective cognitive assessment before discharge because only 63% of the patients over 65 y.o. who wished to drive could return to driving.

A case with bilateral thalamic infarction manifesting mutism - Cerebral blood flow and neural fibers evaluation.

BACKGROUND: Acute bilateral thalamic infarction is rare, and occlusion of the artery of percheron (AOP) may be one of its reasons. AOP occlusion results in an acute disturbance of consciousness, but mutism due to AOP occlusion is rare. We described a mutism patient with bilateral thalamic infarction presumably due to AOP occlusion. We also performed cerebral blood flow (CBF) evaluation by N-isopropyl-p-[123I]-iodoamphetamine single-photon emission computed tomography (123I-IMP-SPECT) as well as neural fiber evaluation by diffusion tensor tractography, discussing the mechanism of mutism. CASE DESCRIPTION: A 92-year-old woman presented a gradual deterioration of consciousness. Diffusion-weighted images revealed high-intensity areas at the bilateral thalami, and we diagnosed AOP occlusion. We administered a recombinant tissue plasminogen activator. On day 14, her Glasgow Coma Scale score was 11 (E4V1M6), and she did not present any apparent paresis. She was mute but cognitively alert, although she could communicate by nodding or facial expression. 123I-IMP-SPECT showed CBF increase in the bilateral cerebellum and CBF decrease in the infarcted bilateral thalami and frontal lobes. Diffusion tensor tractography revealed the bilateral dentatothalamo-cortical tracts (DTCs). However, the tracts terminated at the parieto-occipital cortex, but not at the frontal cortex. She still had mutism on day 30. CONCLUSION: We reported the bilateral thalamic infarction patient presumably due to AOP occlusion, who presented mutism. The discontinuity of the bilateral DTCs resulted in her mutism, and our results supported the hypothesis that the cerebellum plays a significant role in uttering, associated with the bilateral DTCs.

Prediction Model of Deep Learning for Ambulance Transports in Kesennuma City by Meteorological Data.

PURPOSE: With the aging population in Japan, the prediction of ambulance transports is needed to save the limited medical resources. Some meteorological factors were risks of ambulance transports, but it is difficult to predict in a classically statistical way because Japan has 4 seasons. We tried to make prediction models for ambulance transports using the deep learning (DL) framework, Prediction One (Sony Network Communications Inc., Tokyo, Japan), with the meteorological and calendarial variables. MATERIALS AND METHODS: We retrospectively investigated the daily ambulance transports and meteorological data between 2017 and 2019. First, to confirm their association, we performed classically statistical analysis. Second, to test the DL framework's utility for ambulance transports prediction, we made 3 prediction models for daily ambulance transports (total daily ambulance transports more than 5 or not, cardiopulmonary arrest (CPA), and trauma) using meteorological and calendarial factors and evaluated their accuracies by internal cross-validation. RESULTS: During the 1095 days of 3 years, the total ambulance transports were 5948, including 240 CPAs and 337 traumas. Cardiogenic CPA accounted for 72.3%, according to the Utstein classification. The relation between ambulance transports and meteorological parameters by polynomial curves were statistically obtained, but their r2s were small. On the other hand, all DL-based prediction models obtained satisfactory accuracies in the internal cross-validation. The areas under the curves obtained from each model were all over 0.947. CONCLUSION: We could statistically make polynomial curves between the meteorological variables and the number of ambulance transport. We also preliminarily made DL-based prediction models. The DL-based prediction for daily ambulance transports would be used in the future, leading to solving the lack of medical resources in Japan.

Preliminary development of a prediction model for daily stroke occurrences based on meteorological and calendar information using deep learning framework (Prediction One; Sony Network Communications Inc., Japan).

BACKGROUND: Chronologically meteorological and calendar factors were risks of stroke occurrence. However, the prediction of stroke occurrences is difficult depending on only meteorological and calendar factors. We tried to make prediction models for stroke occurrences using deep learning (DL) software, Prediction One (Sony Network Communications Inc., Tokyo, Japan), with those variables. METHODS: We retrospectively investigated the daily stroke occurrences between 2017 and 2019. We used Prediction One software to make the prediction models for daily stroke occurrences (present or absent) using 221 chronologically meteorological and calendar factors. We made a prediction models from the 3-year dataset and evaluated their accuracies using the internal cross-validation. Areas under the curves (AUCs) of receiver operating characteristic curves were used as accuracies. RESULTS: The 371 cerebral infarction (CI), 184 intracerebral hemorrhage (ICH), and 53 subarachnoid hemorrhage patients were included in the study. The AUCs of the several DL-based prediction models for all stroke occurrences were 0.532-0.757. Those for CI were 0.600-0.782. Those for ICH were 0.714-0.988. CONCLUSION: Our preliminary results suggested a probability of the DL-based prediction models for stroke occurrence only by meteorological and calendar factors. In the future, by synchronizing a variety of medical information among the electronic medical records and personal smartphones as well as integrating the physical activities or meteorological conditions in real time, the prediction of stroke occurrence could be performed with high accuracy, to save medical resources, to have patients care for themselves, and to perform efficient medicine.

Where to make burr hole for endoscopic hematoma removal against intracerebral hemorrhage at the basal ganglia to increase the hematoma removal rate - Comparison between trans-forehead and along-the-long-axis approaches.

BACKGROUND: Endoscopic hematoma removal is performed to treat intracerebral hemorrhage (ICH) at the basal ganglia. In our hospital, young neurosurgical trainees perform it for the only 1st to the 3rd time. We perform a "trans-forehead approach" and hypothesized that our technique would contribute to higher hematoma removal rate and easiness despite their inexperience. We compared our dataset with an open dataset with along-the-long-axis approaches using pre- and intraoperative neuronavigation by well-trained neurosurgeons and tested the utility of our trans-forehead approach. METHODS: We retrospectively investigated our 17 consecutive patients with hypertensive ICH who underwent endoscopic hematoma removal using the trans-forehead approach. We obtained the open dataset and compared our data with the 12 patients from the open dataset using the inverse probability weighting method. Operative time, hematoma removal rate, postoperative hematoma volume, Glasgow Coma Scale (GCS) on day 7, and modified Rankin Scale (mRS) at 6 months were assessed as outcomes. RESULTS: The median age was 68 (interquartile range; 58-78) years. Median postoperative hematoma volume, removal rate, operative time, GCS on day 7, and mRS at 6 months were 9 (2-24) mL, 90 (79-98)%, 53 (41-80) min, 13 (12-13), and 4 (2-5), respectively. The weighted generalized estimating equations revealed that operative time was shorter in the along-the-long-axis group, but other items were not significantly different between the two approaches. CONCLUSION: The hematoma removal rate of endoscopic hematoma removal with the trans-forehead approach by young trainees was not different from that of the along-the-long-axis approach by well-trained neurosurgeons using neuronavigation.

Unilateral posterior reversible encephalopathy syndrome characterized with a long and gradually exacerbating course over 3 years and that presented propofol infusion syndrome - A case report.

BACKGROUND: Posterior reversible encephalopathy syndrome (PRES) is characterized by acute neurological symptoms and vasogenic edema, and most patients wholly recover. We report a unilateral PRES patient characterized by a gradual onset followed by propofol infusion syndrome (PRIS) due to general anesthesia therapy. CASE DESCRIPTION: A 32-year-old woman had ovarian dysfunction treated by Kaufmann's treatment for 17 years. Three years ago, she developed seizures, and photophobia and myoclonus sometimes occurred. This time, she had strong photophobia and nausea for 3 months and then developed tonic-clonic seizures for 3 min. Her blood pressure and laboratory test on admission were all within normal limits. She presented no neurological deficits at admission, but the T2-weighted image (T2WI) showed a high-intensity area (HIA), and arterial spin labeling (ASL) image described cerebral blood flow (CBF) increase in the left parieto-occipital region. We diagnosed PRES and started anticonvulsants, antihypertensive, and steroid pulse therapy. However, her aphasia and neuroimaging findings worsened, so we started general anesthesia treatment with propofol on day 29. On day 32, she suddenly developed multiple organ dysfunctions due to PRIS. After intensive care with other sedatives over 2 months, the systemic status and neurological symptoms gradually improved almost as before the onset. On day 90, HIA in the T2WI in the lesion became small, and CBF was severely downregulated in the ASL image. CONCLUSION: Unilateral PRES's pathophysiology and the association with the female hormone remain unknown. Some patients undergo gradual onset and long-term courses, and we should care for PRIS during PRES treatment.

Usefulness of 3 Tesla Ultrashort Echo Time Magnetic Resonance Angiography (UTE-MRA, SILENT-MRA) for Evaluation of the Mother Vessel after Cerebral Aneurysm Clipping: Case Series of 19 Patients.

It is important to assess the cerebral arteries near the clip after cerebral aneurysm clipping. Contrast-enhanced computed tomography angiography has side effects of contrast medium and radiation exposure. Time-of-flight magnetic resonance angiography (TOF-MRA) is a fast and non-invasive method, but clip-induced artifact limits the assessment around the clip. Recently, 3 tesla MRA with ultrashort echo time called SILENT MRA (GE Healthcare Life Sciences, UK) has been reported to have the potential to overcome these disadvantages. We herein present consecutive 19 cerebral aneurysm patients treated by clipping and evaluated using SILENT MRA. The 19 patients (15 women and 4 men) underwent TOF-MRA and SILENT MRA during the same scan session. Two neurosurgeons independently assessed the visibility of the mother vessel at the clipping site in TOF-MRA and SILENT MRA. We also investigated the factors related to visibility in SILENT MRA. All patients' mother vessels were not described in TOF-MRA, and that of 16 patients (84%) were described in SILENT MRA. Overall agreement was 100% in the two neurosurgeons, and the fixed marginal kappa = 1.00 (95% CI: 0.36-1.00). Univariate analysis revealed that larger aneurysm dome and long clip blade length contributed to the visibility of the mother vessel in SILENT MRA. (p = 0.023, 0.007, each). In conclusion, SILENT MRA can be applied for the assessment of the arteries and aneurysm neck remnants near the clip. Using clips with long blade and ligation with its tip would be related to the visibility of the mother vessels in SILENT MRA.

Endoscopically Treated Subacute Subdural Hematoma Presenting Postoperative Cerebral Hyperperfusion Syndrome: Chronological Changes of Cerebral Blood Flow on Arterial Spin Labeling and Subcortical Low Intensity on Fluid-attenuated Inversion Recovery Images.

Subacute subdural hematoma (SASDH) is a neurotraumatic entity. There are few reports of chronological changes of cerebral blood flow (CBF) on arterial spin labeling (ASL) and subcortical low intensity (SCLI) on fluid-attenuated inversion recovery (FLAIR) images of magnetic resonance imaging (MRI) observations from the injury onset, deterioration, to post-surgery. We reported a SASDH patient presenting postoperative cerebral hyperperfusion (CHP) syndrome with chronological changes of those findings. An 85-year-old woman fell and presented right ASDH. She was treated conservatively due to no neurological deficits. On day 3, ASL image revealed increased CBF against brain compression. On day 7, the CBF was normalized on ASL image, but SCLI was confirmed. On day 14, SCLI was strengthened. Then she developed left hemiparesis due to brain compression by SASDH. Considering age and comorbidities, we performed endoscopic hematoma removal under local anesthesia, and her neurological deficits improved after the surgery. On postoperative day 1, she newly presented left upper limb paresis. MRI revealed increased CBF and enhanced SCLI. We diagnosed CHP syndrome, and antihypertensive treatment improved the symptoms gradually. However, SCLI had been consistently observed, and CBF easily changed depending on the blood pressure, suggesting dysfunction of the CBF autoregulation. We showed the endoscopically treated SASDH patient with CBF's chronological changes on ASL images and SCLI on FLAIR images. Long-time brain compression would lead to dysfunction of the CBF autoregulation, and we should be careful about CHP syndrome after the endoscopic surgery for SASDH.

Preliminary development of a deep learning-based automated primary headache diagnosis model using Japanese natural language processing of medical questionnaire.

BACKGROUND: Primary headaches are widespread and costly public health problems. However, there are insufficient medical resources for their treatment in Japan due to two reasons. First, the numbers of headache specialists and clinics remain insufficient. Second, neurologists and neurosurgeons mainly treat headaches in Japan. However, they mainly work as general stroke neurologists, so they cannot focus on primary headache treatment. To solve these problems, we preliminarily developed a deep learning (DL)-based automated diagnosis model from patients' Japanese unstructured sentences in the medical questionnaire using a DL framework. We hypothesized that the model would reduce the time and burden on both doctors and patients and improve their quality of life. METHODS: We retrospectively investigated our primary headache database and developed a diagnosis model using the DL framework (Prediction One, Sony Network Communications Inc., Japan). We used age, sex, date, and embedding layer made by the medical questionnaire's natural language processing (NLP). RESULTS: Eight hundred and forty-eight primary headache patients (495 women and 353 men) are included. The median (interquartile range) age was 59 (40-74). Migraine accounted for 46%, tension-type headache for 47%, trigeminal autonomic cephalalgias for 5%, and other primary headache disorders for 2%. The accuracy, mean precision, mean recall, and mean F value of the developed diagnosis model were 0.7759, 0.8537, 0.6086, and 0.6353, which were satisfactory. CONCLUSION: The DL-based diagnosis model for primary headaches using the raw medical questionnaire's Japanese NLP would be useful in performing efficient medical practice after ruling out the secondary headaches.