金尾 健人

Abstract Background Deep-learning models for prostate cancer detection often require large datasets, which can be challenging to obtain and may lead to domain shift issues in various clinical settings. Purpose This study aimed to develop a deep-learning model for prostate cancer detection on magnetic resonance images using few-shot learning and compare its performance with radiologists. Materials and Methods This retrospective study used 99 cases (80 positive, 19 negative) of confirmed prostate cancer, diagnosed through needle biopsy from 2017 to 2022, with 20 cases for training, 5 for validation, and 74 for testing. The 2D transformer model was trained on T2-weighted, diffusion-weighted, and apparent diffusion coefficient map images. Model predictions were compared between the two radiologists using the Matthews correlation coefficient (MCC) and F1 score, and the bootstrap method was used to calculate 95% confidence intervals (CIs). Results Seventy-four patients (mean age, 71 years ± 8; 60 men) were included in the test set. The model achieved an MCC of 0.297 (95% CI: 0.095–0.474) and F1 score of 0.707 (95% CI: 0.598–0.847). Radiologist 1 had an MCC of 0.276 (95% CI: 0.054–0.484) and an F1 score of 0.741 (95% CI: 0.632–0.832), while Radiologist 2 had an MCC of 0.504 (95% CI: 0.289–0.703) and an F1 score of 0.871 (95% CI: 0.800–0.931). The performance of the model was not significantly different from that of Radiologist 1 (MCC difference: 0.021, 95% CI: −0.270–0.306; F1 score difference: −0.034, 95% CI: −0.153–0.078), but was lower than that of Radiologist 2 (F1 difference: −0.16, 95% CI: −0.287– - 0.061). Conclusion A deep-learning model trained on only 20 cases achieved a performance comparable to one radiologist in detecting prostate cancer on magnetic resonance images, demonstrating the potential of few-shot learning in addressing domain shift challenges. Key Results A deep learning model for prostate cancer detection on MRI was developed using only 20 training cases. The model achieved performance comparable to one radiologist (MCC: 0.297 vs 0.276) but lower than another (F1: 0.707 vs 0.871). Few-shot learning demonstrated potential for addressing domain shift challenges in medical imaging AI. Summary Statement Few-shot learning enables development of prostate cancer detection models on MRI with performance comparable to radiologists, using minimal training data.

BACKGROUND: In Japan, the authorized period (2-4 h) between oral administration of 5-aminolevulinic acid hydrochloride (5-ALA) and transurethral resection for non-muscle invasive bladder cancer (NMIBC) may restrict photodynamic diagnosis (PDD) usage. Therefore, this prospective, single-arm, phase III study aimed to evaluate the diagnostic accuracy and safety of PDD at an extended administration period (4-8 h). METHODS: From January 2022 to May 2023, 161 patients with NMIBC were enrolled from eight hospitals. The primary endpoint was the blue light (BL) sensitivity of pathologically positive biopsies. The secondary endpoints were a comparison of the specificity and positive and negative prediction rates under BL and white light (WL) conditions. RESULTS: A total of 1242 specimens comprising 337 histological NMIBC specimens were analyzed. BL-sensitivity was 95.3%. Its lower limit of 95% confidence interval (92.4-97.3%) exceeded the threshold (70%) of non-inferiority to authorized usage. Sensitivity and specificity were significantly higher and lower for BL than those for WL (95.3% vs. 61.1%, P < 0.001; 52.7% vs. 95.2%, P < 0.001), respectively. The positive and negative predictive rates were significantly lower and higher for BL than those for WL (42.9% vs. 82.7%, P < 0.001; 96.8% vs. 86.8%, P < 0.001), respectively. Of the 145 patients receiving 5-ALA, 136 (93.8%) and 75 (51.7%) experienced 377 adverse events and 95 adverse reactions, respectively, most of which were grade 1 or 2. CONCLUSION: For extended period, the efficacy of PDD for NMIBC was similar to that of authorized period, in terms of higher sensitivity and lower specificity compared with WL, and the safety was acceptable.

UNLABELLED: A 71-year-old man with bone metastasis of hormone-sensitive prostate cancer was treated with androgen deprivation therapy and apalutamide. Radium-223 and radiation therapy were administered after it become castration resistant. Although prostate-specific antigen levels remained low, multiple subcutaneous metastases of neuroendocrine prostate cancer were observed. A review of the pre-treatment prostate needle biopsy revealed a small component with features suggestive of neuroendocrine differentiation. Phosphatase and tensine homolog loss and tumor protein p53 overexpression were observed, confirming the diagnosis of aggressive variant prostate cancer. Platinum-based chemotherapy was administered; however, the patient died 28 months after diagnosis. In this case, if the diagnosis of aggressive variant prostate cancer had been made at an earlier time by biopsy specimens, there might have been a possibility to improve the prognosis by the earlier introduction of the platinum-based regimen. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s13691-024-00673-7.

Choroidal metastasis originating from renal cell carcinomas (RCCs) is rare. To the best of our knowledge, 31 cases of choroidal metastasis from RCC have been reported in the English literature as of January 31, 2024. Nevertheless, physicians need to be vigilant in recognizing this condition, as its progression impacts the quality of life (QOL) of affected patients. In Case 1, a 60-year-old male with a medical history of papillary RCC experienced a deterioration in visual acuity (VA) and was diagnosed with solitary choroidal metastasis. Subsequently, multiple metastases were identified, prompting the initiation of a combination therapy regimen consisting of pembrolizumab plus axitinib. Despite treatment, progression of choroidal metastasis and a further decline in VA were observed. The patient underwent stereotactic radiotherapy and experienced complete resolution of the choroidal metastasis, accompanied by a slight improvement in VA. In Case 2, a 76-year-old man presented with a renal tumor accompanied by lung metastases. He underwent nephrectomy, and the histological diagnosis was papillary RCC. We initiated combination therapy consisting of nivolumab plus cabozantinib. The patient experienced a decrease in VA during treatment. We identified extensive fine metastases scattered throughout the bilateral choroid. We administered axitinib, but the patient experienced bilateral blindness. Given the absence of established therapy for choroidal metastasis, it is crucial to maintain flexibility in treatment selection. Local or systemic approaches should be used as deemed appropriate for each individual case.