Kyoto University Hospital Brain Tumor Surgery Error: Healthy Tissue Removed, Patient on Ventilator

On August 7, 2026, Kyoto University Hospital (京都大学医学部附属病院) held a press conference to admit a serious medical accident during a brain tumor removal surgery in late July. A woman in her 50s underwent a roughly 10-hour craniotomy for a benign deep brain tumor. The surgeon mistakenly removed healthy tissue instead of the tumor, damaging part of the brainstem where respiratory centers are located. The patient cannot breathe spontaneously and remains on a ventilator; the tumor was left in place.

Hospital director Koichi Takaori apologized: “We offer our deepest apologies. We will do our utmost for treatment while investigating causes and implementing effective prevention measures.”

Editor’s note: This article is compiled from Sankei Shimbun, Yomiuri Shimbun, Kyodo News, MBS, TBS, TV Asahi (ANN), Kyoto Shimbun, and hospital statements reported August 7–8, 2026. Medical details reflect public briefings; the investigation is ongoing.

Patient background and why surgery was planned

According to the hospital:

Item	Detail
Patient	Woman in her 50s
Symptoms	Dizziness and unsteadiness for several years; could manage daily life
Diagnosis	Benign tumor (~3 cm) in the cerebellopontine angle — deep region adjacent to the cerebellum and brainstem
Indication	Tumor showed growth; surgery chosen to prevent worsening symptoms
Surgery date	Late July 2026
Procedure	Craniotomy for tumor resection; ~10 hours
Team	Three physicians in the operating room, including the lead surgeon

The cerebellopontine angle is a high-risk area: critical nerves pass nearby, and errors can affect breathing, movement, and consciousness.

What went wrong: timeline of the accident

During surgery

1. A male neurosurgeon in his 40s — with 20+ years in the field and 100+ similar procedures — led the operation.
2. Tissue thought to be tumor was sent for intraoperative frozen-section pathology (術中迅速病理).
3. First result: “Tumor not clearly identified” (腫瘍は明らかではない).
4. The surgeon concluded he had sampled tumor edge tissue where tumor cells might not appear, and continued resection.
5. Second pathology: Same negative finding.
6. Relying on visual judgment and experience, the surgeon believed the tumor was fully removed and ended the case.

After surgery

1. The patient did not regain consciousness or spontaneous breathing after anesthesia.
2. MRI was performed.
3. Findings:
   • The tumor remained in the brain.
   • Healthy tissue had been removed — including damage to the brainstem (呼吸中枢 / respiratory center).
4. Current status (as of the August 7 briefing):
   • On mechanical ventilation; cannot breathe on her own.
   • Limbs not moving; responds to questions with eye opening/closing.
   • Recovery prognosis unclear; treatment continues.

Neurosurgery department chief Arahira Yoshiteru (荒川芳輝) told reporters the core problem was the surgeon’s judgment error: even when pathology twice said “not tumor,” the doctor assumed tumor based on gross appearance — something he said happens occasionally in practice, but should not override objective results. When asked if overconfidence from experience (“慢心”) played a role, he answered: “I think so.” When asked if the surgeon misunderstood and continued, he said: “Yes.”

Patient safety director Matsumura Yumi called it an “unexpected, major medical accident” — because the team had been operating on the wrong anatomical target from the start.

Root causes (as publicly discussed)

Factor	Explanation
Misidentification of surgical target	Resection focused on non-tumor tissue while the actual tumor was left behind
Override of pathology	Two intraoperative reports contradicted the surgeon’s assumption; surgery continued
Cognitive bias / overconfidence	Department chief cited visual fixation and veteran “complacency”
Insufficient team challenge	Three doctors were present, but no effective stop when pathology conflicted with the plan
High-complexity anatomy	Deep location near brainstem magnified consequences of target error

The hospital has preserved full surgical video and says it can reconstruct who moved where during the operation.

Hospital response and prevention measures

Immediate actions

• Apology to patient and family; commitment to best ongoing care.
• Reported to national authorities, Kyoto prefectural/city bodies, and Kyoto Prefectural Police (Yomiuri).
• Established an accident investigation committee including external members.
• Suspended normal disclosure of surgeon identity in some reports; internal review ongoing.

Stated prevention direction

At the briefing, the hospital said that for brain tumor resections, it is now strengthening multi-person verification — implying that critical steps (target confirmation, pathology discordance, extent of resection) should not rest on one surgeon’s solo judgment.

Director Takaori pledged “effective, practical recurrence-prevention measures” after the investigation.

What was not yet detailed publicly

As of August 8 reporting, the hospital had not released a full written protocol listing specific new rules (e.g., mandatory pause when pathology negates tumor, mandatory neuronavigation cross-check, second-surgeon sign-off). Those may follow the external committee review.

How other major hospitals and systems approach this risk

Kyoto University Hospital performs about 156 brain tumor surgeries per year — a high-volume center. The accident nonetheless shows that volume and experience do not eliminate catastrophic error when safeguards fail.

Intraoperative pathology: standard practice and limits

Frozen-section pathology during brain surgery is routine at university hospitals. It helps confirm tumor type and resection margins — but it is not infallible:

• Small samples, freezing artifacts, and borderline tissue can yield false negatives or ambiguous reads.
• Japanese pathology guidance stresses cytology plus histology, rapid immunostaining when needed, and pre-op conferences between surgeons and pathologists.
• When results conflict with clinical suspicion, best practice is to pause, re-sample, re-image, or consult — not assume the pathologist missed tumor at the edge.

The Kyoto case is unusual because pathology twice denied tumor, yet resection continued — suggesting a process breakdown, not merely a difficult diagnosis.

Technology-assisted safeguards (examples at other centers)

Tool	Purpose	Example institutions / research
Neuronavigation	Real-time 3D localization of tumor vs. eloquent brain	Standard at major neurosurgery centers
Intraoperative MRI	Confirm residual tumor before closing	High-end brain tumor programs
Motor evoked potential (MEP) / mapping	Protect motor pathways	Tokyo Women’s Medical University and others
Awake surgery	Language/motor testing during resection	Used in hundreds of cases at experienced centers
Intraoperative flow cytometry (iFC)	Objective DNA-based tumor vs. normal cell distinction in ~10 min	Clinical research programs in Japan
Integrated diagnosis “i-ID”	Combines rapid pathology + molecular markers in ~90 min	Yokohama City University (reported ~95–97% concordance with final diagnosis)
Fluorescence-guided surgery	Visualize tumor tissue (e.g., 5-ALA)	Glioma programs worldwide

These tools aim to reduce reliance on a single surgeon’s visual guess — especially near the brainstem.

Team and system safeguards

Japan’s JONA Perioperative Safety Checklist (2024) and the WHO Surgical Safety Checklist require team confirmation at sign-in, time-out, and sign-out — including specimen labeling and critical step review.

WHO guidance explicitly treats pathology labels and specimen identity as high-risk steps. JONA Phase 3 checks before closing include procedure confirmation, instrument counts, and specimen verification.

What checklists cannot do alone: force a surgeon to stop when pathology contradicts their mental model. That requires culture — psychological safety for nurses and junior doctors to speak up, and hard stops when objective tests fail twice.

Reporting and learning systems

Japan’s medical accident reporting framework encourages hospitals to disclose serious incidents, involve external review, and share lessons. High-profile university hospital cases often drive national protocol updates — but implementation lag remains a chronic issue.

Why this case matters beyond Japan

1. Information asymmetry: Patients cannot verify which millimeters of brain are being cut. They delegate trust to a team they cannot supervise.
2. Autonomy vs. accountability: Surgeons need real-time decision power in the OR — but unchecked autonomy plus cognitive bias can be lethal.
3. Technology’s promise and limits: Navigation, AI-assisted cytometry, and integrated molecular diagnosis can objectify decisions — yet someone must still honor a negative pathology report.
4. The AI question: AI may eventually assist targeting and risk scoring, but legal, ethical, and emergency accountability still center on humans today.

---

Discussion

This patient entered surgery to prevent dizziness from a benign tumor. She left with brainstem injury and a ventilator. The investigation continues — but the structural questions are already here.

1. How can we change the asymmetric relationship between patients and operating surgeons?

Should patients have the right to mandated second opinions on surgical plans, recorded consent videos explaining risks, or independent safety officers in the OR? What would actually work without paralyzing emergency care?

2. How can we effectively prevent — and correct — an individual surgeon’s subjective bias?

Mandatory pause rules when pathology contradicts the surgeon? Anonymous near-miss reporting? Rotation off high-risk cases after override events? Peer review of surgical video?

3. Would you entrust your life to a doctor with full autonomous decision-making power in the operating room?

If you already do implicitly by signing consent forms, what additional safeguards would you require before doing so again?

4. If AI could assist or even lead intraoperative decisions, would you choose the doctor or the AI?

Under what conditions — only as a navigation aid? As a veto when pathology and imaging disagree? Full autonomy? Who bears liability when the algorithm is wrong?

Share your view: trust, regulation, technology, or something else first?

---

This article is news reporting, not medical advice. For health decisions, consult qualified physicians in your jurisdiction.