Performance of Generative Pretrained Transformer on the National Medical Licensing Examination in Japan

The remarkable performance of ChatGPT, launched in November 2022, has significantly impacted the field of natural language processing, inspiring the application of large language models as supportive tools in clinical practice and research worldwide. Although GPT-3.5 recently scored high on the United States Medical Licensing Examination, its performance on medical licensing examinations of other nations, especially non-English speaking nations, has not been sufficiently evaluated. This study assessed GPT's performance on the National Medical Licensing Examination (NMLE) in Japan and compared it with the actual minimal passing rate for this exam. In particular, the performances of both the GPT-3.5 and GPT-4 models were considered for the comparative analysis. We initially used the GPT models and several prompts for 290 questions without image data from the 116th NMLE (held in February 2022 in Japan) to maximize the performance for delivering correct answers and explanations of the questions. Thereafter, we tested the performance of the best GPT model (GPT-4) with optimized prompts on a dataset of 262 questions without images from the latest 117th NMLE (held in February 2023). The best model with the optimized prompts scored 82.7% for the essential questions and 77.2% for the basic and clinical questions, both of which sufficed the minimum passing scoring rates of 80.0% and 74.6%, respectively. After an exploratory analysis of 56 incorrect answers from the model, we identified the three major factors contributing to the generation of the incorrect answers-insufficient medical knowledge, information on Japan-specific medical system and guidelines, and mathematical errors. In conclusion, GPT-4 with our optimized prompts achieved a minimum passing scoring rate in the latest 117th NMLE in Japan. Beyond its original design of answering examination questions for humans, these artificial intelligence (AI) models can serve as one of the best "sidekicks" for solving problems and addressing the unmet needs in the medical and healthcare fields.

A clinical assessment of portable point-of-care testing for quick cortisol assay during adrenal vein sampling

This study assessed the clinical performance of point-of-care testing (POCT) for quick cortisol assay (QCA) during adrenal vein sampling (AVS) using a newly invented portable quantitative assay instrument. An observational study was conducted prospectively at two centres in Japan. Forty-eight patients with primary aldosteronism considered for adrenalectomy were enrolled in this study and underwent AVS. Three basal adrenal vein samples from each adrenal vein and two from the inferior vena cava were collected sequentially. The cortisol concentration of adrenal vein samples was measured by routine method and QCA. A total of 338 adrenal vein samples were analysed from 250 sites to determine AVS success or failure. The distribution of turnaround time of the QCA for AVS success or failure followed a normal distribution with an average of 20.5 min. A positive correlation between the routine method and QCA was observed regarding cortisol concentration or selectivity index. No significant difference between the two methods was observed regarding the success rate of AVS. Using the routine method as a reference, the sensitivity and specificity of AVS success or failure were 99.1% (210/212) and 81.6% (31/38), respectively. Easy, quick, portable, and precise POCT-QCA demonstrated its compatibility with routine methods regarding clinical performance.

In vivo relationship between bound and free insulin in patients with diabetes having anti-insulin antibodies

Objective

The Scatchard plot of anti-insulin antibodies is curvilinear, indicating heterogeneity in binding sites. However, the relationship between bound insulin (B) and free insulin (F) in patients with anti-insulin antibodies has not yet been elucidated. This study aimed to determine this relationship.

Methods

We studied two insulin-treated patients with diabetes who had high titers of anti-insulin antibodies. The B and F levels were measured using daily blood samples. Assuming that the law of mass action is applicable to the reactions between insulin and anti-insulin antibody forms, we plotted the bound-to-free ratio (B/F) vs. B using patient data. We also performed an equilibrium binding assay in vitro.

Results

Some of the B/F vs. B plots of the daily variation showed an approximately linear relationship, while the Scatchard plots of in vitro data became curvilinear.

Conclusion

Our study suggests that the one-site (high-affinity site) of anti-insulin antibodies accounts, for the most part, for insulin pharmacokinetics within physiological insulin concentrations.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13340-023-00641-1.

Acquisition of Interleukin-6 Production Ability Over Time With Pheochromocytoma

Pheochromocytoma is a tumor of chromaffin cells causing catecholamines overproduction. Interleukin-6 (IL-6), a cytokine, is central to inflammation and immunity. Few studies have reported IL-6-producing pheochromocytoma whose underlying mechanism has not been elucidated. Herein, we present a case of pheochromocytoma whose clinical manifestations changed, and IL-6 levels elevated over time. A 48-year-old woman was referred to our hospital for fever and hepatic dysfunction. Six years prior, a right adrenal tumor was detected during the examination for ovarian teratoma without C-reactive protein (CRP) elevation. Several imaging studies at our hospital showed no abnormalities except for an increase in the size of the adrenal tumor and hepatomegaly. In addition, antibiotics did not improve the fever. Laboratory tests showed elevated levels of CRP with IL-6 elevation. An enlarged adrenal tumor was detected. Administering doxazosin lowered the CRP and IL-6 levels, then IL-6-producing pheochromocytoma was suspected, and adrenalectomy was performed. After surgery, fever and hepatic function were improved, and the CRP and IL-6 levels were normalized. Immunostaining of the resected tissue showed IL-6 focal positivity, which meant the phenotype of tumor cells focally changed their phenotypes over time. IL-6-producing pheochromocytoma should be considered in patients with adrenal tumors and fever of unknown origin.

Using telemedicine to manage a patient with a hypertensive emergency due to pheochromocytoma

Summary

The COVID-19 pandemic has led to the emergence of telemedicine on a global scale. In endocrinology, telemedicine has mainly been used in relation to chronic diseases, including diabetes. Herein, we report the case of an 18-year-old woman with a hypertensive emergency due to a pheochromocytoma who was quickly diagnosed and treated using telemedicine. The patient was referred to a cardiovascular hospital because of fatigue and sweating that did not improve with carvedilol. She had fluctuating blood pressure and tachycardia. Subsequently, since her thyroid function was normal, endocrine hypertension not due to thyroid dysfunction was suspected; a case consultation was made by phone to our clinic. Plain computed tomography (CT) was recommended owing to the high possibility of a pheochromocytoma; the CT scan showed an adrenal tumor with a 30 mm diameter. To assess her condition, endocrinologists, together with the attending doctor, interviewed her and her family directly using an online tool to obtain detailed information. We thus determined that she was at risk of a pheochromocytoma crisis. She was transferred to our hospital immediately for treatment, was diagnosed with pheochromocytoma, and underwent surgery. Telemedicine, especially involving doctor-to-patient with doctor consultations, can be effective in treating rare and emergent medical conditions such as pheochromocytoma crisis.

Learning points

Telemedicine can be used for chronic diseases and emergency conditions. Online doctor-to-patient with doctor (D-to-P with D) consultations are useful when the expert opinion of a highly specialized physician present in a different geographical location is required. Telemedicine, especially D-to-P with D online consultations, can be effectively used for the diagnosis of rare and emergent medical conditions, such as pheochromocytoma crisis.

Impact of Conducting Adrenal Venous Sampling in the Morning Versus Afternoon in Primary Aldosteronism

Context

Adrenal venous sampling (AVS) is the gold standard technique for subtype differentiation of primary aldosteronism (PA) and to obtain aldosterone and cortisol measurements; however, their secretion patterns show fluctuations during the day.

Objective

We aimed to examine the effects of AVS timing on AVS results.

Methods

This multicenter, retrospective, observational study included a total of 753 patients who were diagnosed with PA and underwent AVS in 4 centers in Japan. Among them, 504 and 249 patients underwent AVS in the morning (AM-AVS) and in the afternoon (PM-AVS), respectively. The outcome measures were the impact of AVS timing and hormone fluctuations in a day on AVS results.

Results

There were no differences in the success rate of AVS, diagnostic rate of disease type, or frequency of discrepancy in PA subtypes between the AM-AVS and PM-AVS groups. Regarding patients with unilateral PA, aldosterone concentrations in adrenal venous blood did not differ between the 2 groups on the dominant or nondominant side. Conversely, regarding patients with bilateral PA, aldosterone concentrations in adrenal venous blood were significantly higher in the AM-AVS than in the PM-AVS group.

Conclusions

The timing of AVS did not seem to have a significant impact on subtype diagnosis. The aldosterone levels in adrenal venous blood were significantly higher in patients with bilateral PA in the AM-AVS group, but there was no such difference between patients with unilateral PA in the AM-AVS and PM-AVS groups. Each subtype may have a different hormone secretion pattern in a day.

Recurrence of Primary Aldosteronism After Surgery in Aldosterone-producing Adenoma With KCNJ5 Gene Mutation

Primary aldosteronism (PA) is a major cause of secondary hypertension. Aldosterone-producing adenoma (APA) is a subtype of PA, and adrenalectomy is the gold-standard treatment. Recently, a high prevalence of the KCNJ5 gene mutation has been reported in APA, particularly in Japan. Herein, we present 2 extremely rare cases of PA recurrence more than 10 years after adrenalectomy for APA. In the first case, a 52-year-old woman was examined for hypertension 22 years after total adrenalectomy of the right adrenal gland. Recurrent PA was diagnosed based on high aldosterone-renin-ratio (ARR), identification of left adrenal gland tumor by computed tomography (CT), and a confirmatory test. In the second case, a 65-year-old man was examined for hypertension 17 years after total adrenalectomy of the left adrenal gland. He had maintained his blood pressure using medication since the onset of hypertension 4 years after the surgery. A year later, a high ARR was observed. PA recurrence was determined by a right adrenal gland tumor noted on CT and a confirmatory test. Somatic mutations in KCNJ5 were detected in the resected tissues in both cases. We recommend careful follow-ups after adrenalectomy in APA cases, especially in those with a KCNJ5 gene mutation.