Biopsies are the most reliable tool medicine has for diagnosing cancer and many other diseases, but they are not infallible. When the same biopsy slides are reviewed by a second pathologist, the original diagnosis is confirmed roughly 90% of the time for straightforward cases, yet agreement drops sharply for borderline or ambiguous findings. Errors can creep in at every stage, from the moment the needle enters tissue to the second a pathologist renders a verdict under the microscope. Understanding where and why biopsy results go wrong helps you know when to trust a result, when to push for more information, and when a second opinion could genuinely change your care.
How Often Pathologists Agree With Each Other
The single most-studied measure of biopsy accuracy is what happens when a second pathologist reviews the same slides. In breast pathology, a large decision-analysis study found that without any second opinion, about 92% of biopsies received a concordant diagnosis. When all biopsies were reviewed by a second pathologist, that figure climbed to roughly 95%, with under-treatment falling to its lowest rate.1PubMed Central. Second opinion strategies in breast pathology: a decision analysis addressing over-treatment, under-treatment, and care costs Those numbers sound reassuring until you consider how many biopsies are performed each year. Even a small percentage of discordance translates to a large number of people whose treatment plan could change.
A study at a referral cancer center that specifically re-examined cases sent for second opinions found a much higher mismatch rate: about 41% of cases showed some discrepancy, with roughly 17% having major discrepancies that could alter the diagnosis or treatment approach.2PubMed. Pathology Reports: Discrepancy Patterns of Second Opinions in a Referral Cancer Center That figure is higher than average because these were cases already flagged as difficult or unusual enough to warrant referral. In a separate review of breast pathology second opinions across nearly 2,000 cases, clinically significant discrepancies that actually changed patient care appeared in about 11.5% of reviews.3PubMed. Breast pathology second review identifies clinically significant discrepancies in over 10% of patients The takeaway is that most biopsies get the right answer the first time, but a meaningful minority do not, and the harder the case, the higher the odds of disagreement.
Why Some Diagnoses Are Harder Than Others
Not all biopsy diagnoses carry the same level of certainty. Clear-cut cancers and obviously benign tissue are the easy calls. The trouble zones are the in-between categories: pre-cancerous changes, low-grade abnormalities, and lesions that sit on the border between benign and malignant. In breast pathology, a landmark study found that pathologists agreed with the consensus reference diagnosis on fewer than half of atypia cases, with a concordance rate of just 48%.4JAMA. Diagnostic Concordance Among Pathologists Interpreting Breast Biopsy Specimens Atypia matters because it often triggers surveillance or even surgery, yet the line between “atypical but not cancer” and “early cancer” is genuinely blurry under the microscope. In one study of cases where pathologists had to distinguish atypical ductal hyperplasia from ductal carcinoma in situ, all five reviewing pathologists agreed on the diagnosis in only 30% of cases.5PubMed Central. Atypical Ductal Hyperplasia Bordering on Ductal Carcinoma In Situ: Interobserver Variability and Outcomes in 105 Cases
Melanoma presents a similar challenge. Among expert dermatopathologists reviewing clinically difficult pigmented skin lesions, agreement was strong for clear-cut melanoma versus non-melanoma. But for borderline lesions, particularly in younger patients and those with atypical mole syndrome, agreement dropped considerably. In patients with atypical mole syndrome, inter-observer agreement fell to near-chance levels.6Dermatology. Agreement of Dermatopathologists in the Evaluation of Clinically Difficult Melanocytic Lesions: How Golden Is the ‘Gold Standard’? A larger study of pathologists diagnosing melanocytic tumors confirmed that overall agreement was substantial for straightforward melanomas and common moles but dropped to only moderate levels for non-invasive (in situ) melanomas.7PubMed Central. Discordance, accuracy and reproducibility study of pathologists’ diagnosis of melanoma and melanocytic tumors
The Sampling Problem
Even a perfectly skilled pathologist can only diagnose what is on the slide. If the needle missed the abnormal area, or if the tissue sample happened to catch a region of inflammation rather than tumor, the result will be misleading regardless of who reads it. Breast cancers are a good example: a tumor can contain not just cancer cells but pockets of scar tissue, dead cells, inflammatory areas, and pre-cancerous changes all mixed together. A core needle biopsy might sample a non-cancerous pocket, producing a false-negative result, or it might catch a pre-cancerous area and miss the invasive cancer next door, leading to an underestimate of the disease.8PubMed Central. False-negative results of breast core needle biopsies – retrospective analysis of 988 biopsies
Liver biopsies highlight sampling error in a different way. In patients with chronic hepatitis C, researchers biopsied both the right and left lobes and had the same pathologist read both samples. About a quarter of patients had a difference of at least one grade of inflammation between the two sides, and a third had a difference of at least one stage of scarring. In roughly 15% of patients, one lobe was read as cirrhosis while the other was read as one step below cirrhosis.9PubMed. Sampling error and intraobserver variation in liver biopsy in patients with chronic HCV infection The pathologist was not making a mistake; the disease was simply more advanced in one part of the liver than the other, and a single needle pass could land in either zone.
How Tissue Handling Can Introduce Errors
Before a pathologist ever looks at a slide, the tissue goes through a chain of steps: surgical removal, preservation in fixative, processing, embedding in wax, slicing into thin sections, and staining. Defects introduced at any of these stages are known as artifacts, and they can distort cells enough to mimic disease or mask it.10PubMed Central. Facts in artifacts A crush artifact from forceps, for instance, can flatten cells so they look abnormally dense, while poor fixation can cause tissue to degrade before it reaches the lab.11PubMed Central. Artefacts: a diagnostic dilemma – a review Quality-control data from pathology labs show that a small but real fraction of specimens arrive without proper fixative, compromising interpretation.12PubMed Central. Quality Measures in Pre-Analytical Phase of Tissue Processing: Understanding Its Value in Histopathology These pre-analytical errors are invisible to the patient but can be the hidden reason a result is wrong.
Needle Type Matters
The two most common biopsy techniques are fine-needle aspiration (FNA), which uses a thin needle to suction out individual cells, and core needle biopsy, which cuts out a small cylinder of intact tissue. Core biopsies preserve the architecture of the tissue, giving the pathologist more information to work with. In breast lesions with ambiguous imaging, core needle biopsy had an overall diagnostic accuracy of about 94%, compared to roughly 80% for FNA.13Journal of Heart Valve Disease. Comparative Diagnostic Accuracy of Core Needle Biopsy versus Fine Needle Aspiration Cytology in Breast Lesions with Atypical Imaging Features For soft-tissue masses in the arms and legs, FNA and core biopsy had similar sensitivity for detecting malignancy (both around 79%), but core biopsy edged ahead in overall accuracy.14PubMed Central. A Comparison of Fine-needle Aspiration, Core Biopsy, and Surgical Biopsy in the Diagnosis of Extremity Soft Tissue Masses
Thyroid nodules illustrate a common real-world scenario: FNA is the standard first step, but it often returns an “indeterminate” result that leaves both doctor and patient in limbo. When a core needle biopsy is used as a follow-up for those indeterminate thyroid cases, it can reclassify up to 98% of them as clearly malignant or clearly benign.15PubMed Central. Core needle biopsy in the management of thyroid nodules with an indeterminate fine-needle aspiration report The trade-off is that core biopsies are slightly more invasive and carry modestly higher complication rates, so not every situation calls for one.
How Imaging Guidance Improves Accuracy
Where the needle goes matters as much as what kind of needle is used. Blind or freehand biopsies are increasingly rare for solid tumors because image-guided techniques, using ultrasound, CT, or MRI to steer the needle in real time, dramatically improve the odds of hitting the right spot. In prostate cancer, MRI-guided biopsy detects clinically significant cancers at a higher rate than the traditional systematic approach, while picking up fewer insignificant cancers that might lead to unnecessary treatment.16PubMed. Image-guided prostate biopsy using magnetic resonance imaging-derived targets: a systematic review
The gains become clearer when methods are combined. In a study of over 800 men, systematic biopsy alone detected clinically significant prostate cancer in 40% of cases, MRI-targeted biopsy alone caught 43%, and the combination of both reached 50%.17British Journal of Radiology. Accuracy of MRI-ultrasound fusion-guided and systematic biopsy of the prostate A meta-analysis confirmed that MRI-ultrasound fusion biopsy detected more clinically significant cancers than systematic biopsy alone.18PubMed Central. Is magnetic resonance/ultrasound fusion prostate biopsy better than systematic prostate biopsy? An updated meta- and trial sequential analysis Fusion imaging, which overlays data from two imaging methods, is one of the most promising advances in tissue sampling and is gradually becoming standard for challenging biopsy targets.19PubMed Central. Recent Advances in Image-Guided Tissue Sampling
When the Result Comes Back “Nondiagnostic”
Sometimes the problem is not a wrong answer but no answer at all. A biopsy may return as “nondiagnostic” or “insufficient for diagnosis,” meaning the pathologist did not have enough tissue to make a call. For kidney biopsies, one national laboratory tracked this over 16 years and found that the rate of inadequate samples climbed from about 2% in 2005 to 14% by 2020, with an overall miss rate of 11%.20PubMed Central. Increasing Incidence of Inadequate Kidney Biopsy Samples Over Time: A 16-Year Retrospective Analysis From a Large National Renal Biopsy Laboratory The reasons likely include changing practice patterns and the growing use of less invasive biopsy techniques that retrieve smaller tissue samples.
A nondiagnostic result should not be treated as reassurance that nothing is wrong. In a study of small kidney masses, about 19% of initial biopsies were nondiagnostic. When pathology was eventually available for those masses (from surgery or repeat biopsy), 73% turned out to be malignant.21PubMed. Outcomes of small renal mass needle core biopsy, nondiagnostic percutaneous biopsy, and the role of repeat biopsy Repeat biopsy successfully provided a diagnosis in most cases. For endometrial biopsies, insufficient samples are especially common in older women, and when worrisome features were present in the initial nondiagnostic sample, 43% of patients turned out to have uterine malignancy on follow-up procedures.22International Journal of Gynecological Pathology. Clinical Outcomes of Patients With Insufficient Sample From Endometrial Biopsy or Curettage The message for patients: if your biopsy comes back as insufficient, ask your doctor about a repeat procedure rather than assuming the results are reassuring.
Liquid Biopsies and Their Limitations
Liquid biopsy, which analyzes fragments of tumor DNA circulating in your blood, has generated excitement as a less invasive alternative to needle biopsies. For metastatic non-small-cell lung cancer, one study found liquid biopsy was roughly 95% to 100% concordant with tissue biopsy for guideline-recommended biomarkers, and a liquid-first approach actually identified actionable mutations in more patients than a tissue-first approach.23PubMed. Liquid Biopsy Versus Tissue Biopsy to Determine Front Line Therapy in Metastatic Non-Small Cell Lung Cancer (NSCLC) However, a direct head-to-head comparison at a single institution painted a more cautious picture. Tissue-based sequencing caught about 95% of clinically relevant mutations, while blood-based sequencing caught only about 53%, and overall concordance between the two was just 59%.24Modern Pathology. Comparison of solid tissue sequencing and liquid biopsy accuracy in identification of clinically relevant gene mutations and rearrangements in lung adenocarcinomas
The practical lesson is that a positive liquid biopsy result is highly trustworthy, since finding tumor DNA in blood almost certainly means the mutation exists. But a negative liquid biopsy is far less reliable, because many tumors do not shed enough DNA into the bloodstream to be detected. Most oncologists currently treat liquid biopsy as a complement to tissue biopsy rather than a replacement, using it when tissue is hard to obtain or when speed matters for treatment decisions.
How AI Is Changing Pathology
Artificial intelligence is increasingly being used as a second set of eyes in the pathology lab. A systematic review and meta-analysis of AI tools in digital pathology found a mean sensitivity of about 96% and specificity of about 93% across a broad range of diagnostic tasks.25npj Digital Medicine. Artificial intelligence in digital pathology: a systematic review and meta-analysis of diagnostic test accuracy These tools are especially useful for repetitive quantification tasks where human eyes tire. In one study, pathologists using an AI tool to score a specific protein marker in breast cancer reduced their scoring error by more than half compared to working without AI, and agreement among pathologists also jumped substantially.26Scientific Reports. AI improves accuracy, agreement and efficiency of pathologists for Ki67 assessments in breast cancer
AI is not replacing pathologists. The current model is closer to a spell-checker: the algorithm flags areas of concern or provides quantitative measurements, and the pathologist makes the final call. This pairing reduces the kind of inter-observer variability described earlier and helps catch subtle patterns that a tired human might overlook.27PubMed Central. Transforming Diagnostics: A Comprehensive Review of Advances in Digital Pathology Access remains uneven, though, with large academic centers adopting these tools faster than community hospitals.
When a Second Opinion Changes Treatment
Given everything above, seeking a second pathology opinion is one of the most practical steps a patient can take, particularly for diagnoses in the gray zone. A retrospective review of 120 oncology cases sent for second opinions found that 42 patients had clinically meaningful changes in their treatment plans. Of those changes, all 42 involved improved expected morbidity, and 11 also had improved expected prognosis. Nine patients shifted from active treatment to observation, and 21 had surgery reduced or eliminated.28PubMed Central. Clinical value of second opinions in oncology: A retrospective review of changes in diagnosis and treatment recommendations For bladder cancer patients referred to a comprehensive cancer center, second opinions led to potential treatment changes in about 15% of cases, including altered surgical recommendations in roughly 7%.29PubMed. Change in Management Based on Pathologic Second Opinion Among Bladder Cancer Patients Presenting to a Comprehensive Cancer Center: Implications for Clinical Practice
The financial picture adds another argument for second opinions. In the same 120-case oncology review, the average cost saving across all patients (including those whose treatment did not change) was over $15,000 per patient, because the revised plans tended to involve less aggressive surgery or medication regimens.30PubMed Central. What Is the Cost Impact of Second Opinions in Oncology? A Retrospective Review Second opinions are most valuable for borderline diagnoses (atypia, low-grade lesions, in situ versus invasive carcinoma), rare tumor types, and any situation where the recommended treatment is irreversible.
Complications and What to Expect Physically
Accuracy is not the only concern patients have. The physical risks of biopsy, while generally low, vary with the technique and the target organ. In a study comparing ultrasound-guided and CT-guided biopsies, overall complication rates including minor events were about 8% for ultrasound-guided and 30% for CT-guided procedures. The difference was driven largely by bleeding and air leakage into the chest cavity during lung biopsies. The vast majority of complications were mild, and no life-threatening events occurred.31PubMed Central. Navigating Biopsy Safety: Complication Rates Under Ultrasound and CT Guidance A meta-analysis focused specifically on CT-guided lung biopsies found higher overall complication rates (about 39% for core biopsy, 24% for FNA), though major complications requiring intervention stayed in the range of 4% to 6%.32PubMed Central. Complication rates of CT-guided transthoracic lung biopsy: meta-analysis If you have high blood pressure, the good news is that it does not appear to meaningfully increase your bleeding risk from a core needle biopsy. A large study of nearly 5,000 biopsies found major hemorrhage rates of about 0.3%, with no significant difference between hypertensive and normotensive patients.33PubMed. Incidence of bleeding complications after percutaneous core needle biopsy in hypertensive patients and comparison to normotensive patients
The Emotional Weight of Waiting
Biopsy accuracy is a scientific question, but for the person sitting in the waiting room, it is deeply personal. Research on women awaiting breast biopsy results found that distress levels were high across the board. Patients who had a family history of breast cancer, had undergone previous biopsies, or used avoidant coping strategies reported the greatest distress. About half of the women overestimated their personal risk of a positive finding.34PubMed. Waiting for a breast biopsy. Psychosocial consequences and coping strategies A longitudinal study that tracked women daily during the wait for results found that anxiety peaked immediately after the biopsy and again in the final days before results were delivered, suggesting that targeted support at those two moments could make the biggest difference.35PubMed. Emotional, cognitive, and physical well-being during the wait for breast biopsy results
Understanding that biopsies, while highly accurate for clear diagnoses, are genuinely uncertain for borderline findings can paradoxically reduce some of that anxiety. If a doctor recommends additional testing or a second opinion, it does not mean the first result was a failure. It means the case falls into a category where even experts disagree, and the responsible next step is to gather more information. Expecting certainty from a single biopsy can set you up for distress when the reality of medicine is often more nuanced than that.
Biopsy Accuracy in Veterinary Medicine
Humans are not the only patients whose biopsies sometimes miss the mark. In veterinary oncology, the same sampling and interpretation challenges apply. A study of pre-treatment biopsies for soft-tissue tumors in dogs found that the biopsy grade matched the final surgical grade only 59% of the time. Of the mismatches, 29% underestimated and 12% overestimated the tumor’s aggressiveness.36PubMed. Diagnostic accuracy of pre-treatment biopsy for grading soft tissue sarcomas in dogs For canine bone lesions, cytology and histology had similar overall accuracy, around 82% to 83%, but correctly identifying the specific tumor type was possible only about half the time with either method.37Journal of Veterinary Internal Medicine. Comparative Assessment of the Accuracy of Cytological and Histologic Biopsies in the Diagnosis of Canine Bone Lesions These findings mirror the human experience: biopsies are good at distinguishing cancer from non-cancer, less good at precise grading and subtyping, and sensitive to where exactly the needle lands. If your veterinarian recommends a repeat biopsy or referral to a specialist, the reasoning is the same as in human medicine.