A previous healthy 61-year-old man visited a local doctor with a chief complaint of a mass on his head. He was referred to the neurosurgery department of our hospital on suspicion of a tumor in the skull. Magnetic resonance imaging of the head was performed by 3 T device (GE DISCOVERY MR750; GE Healthcare) with Gadolinium-based contrast agent, and metastatic tumor was suspected. Blood examination showed a high prostate-specific antigen level of 165.42 ng/mL, and a computed tomography scan detected lytic lesions of pelvic bone adjacent to the abnormally enlarged prostate.. 99mTc-Hydroxymethylene diphosphonate bone scintigraphy revealed multiple bone metastases throughout the body. The patient was further referred to our department on suspicion of multiple bone metastases from prostate cancer. A needle biopsy of the patient’s prostate revealed an adenocarcinoma with a Gleason score of 5 + 4 = 9, which prompted ADT. Elevation of the prostate-specific antigen level and postrenal failure due to urinary retention were observed 1 year after ADT started, and channel transurethral resection of the prostate (TURP) was thus performed. Docetaxel was administered for a total of seven courses because the patient’s tumor was considered to have acquired castration resistance. However, multiple lymph node and bone metastases were exacerbated. Although a total of six courses of cabazitaxel had already been administered, it was discontinued owing to sacral pressure ulcer infection and deterioration of the patient’s general condition. He was subsequently treated with abiraterone, but he died 2 years 6 months after being diagnosed with prostate cancer. We performed genetic sequencing of 160 cancer-related genes (PleSSision-Rapid®) using specimens collected by TURP. RB1 and BRCA2 co-deletion as well as a truncating mutation of TP53 (G244Rfs*19) were detected. Amplification of the androgen receptor (AR) gene was observed with an estimated copy number of 18.3. In addition, a patched 1 (PTCH1) point mutation (p.R441H) was detected as a possibly pathogenic alteration. Immunohistochemical staining was performed using standard protocols. The products of the antibodies and the detailed protocols are shown in the additional file. All stained sections were scanned using a high-resolution digital slide scanner (NanoZoomer-XR C12000; Hamamatsu Photonics, Hamamatsu, Shizuoka, Japan), which consisted of a trilinear sensor camera, a detector of 4096 pixels × 64 lines x 3plates, and a filter that only divides into RGB with a prism. The measured resolution of all microscopy images was 0.23 μm/pixel, which was equivalent to 40 × objective lens. Hematoxylin and eosin staining showed that tumor cells had clear nucleoli and a histology different from that of typical neuroendocrine prostate cancer cells. Complete absence of RB1 and p53 protein expressions (Additional file: Fig. S1A–B) were consistent with genomic findings. AR was stained in 70% of tumor nuclei and PSA was positive in about 30% of tumor cells (Additional file: Fig. S1C–D), whereas neuroendocrine markers were not stained (Additional file: Fig. S1E–G). Immunohistochemical staining of glioma-associated oncogene family zinc finger 1 (GLI1) was performed to confirm whether the hedgehog signal was enhanced by a mutation of PTCH1 in our case. GLI1-positive tumor cells accounted for less than 10% of the total (Additional file: Fig. S1H) and the nuclear stainability of the cells was not stronger than that of Leydig cells used as positive control (data not shown), indicating that the hedgehog signal was not enhanced.