Pathology Department

What our pathology unit covers — and who it is for

Pathology is the specialty that decides what something is. When tissue is removed — a biopsy, a polyp, a lump, a whole organ — the answer to the only question that matters does not come from the operation or the scan. It comes from a pathologist looking at that tissue under a microscope and writing a report. Almost every cancer diagnosis in medicine begins as a sentence in a pathology report.

This is also the specialty patients never meet. The surgeon has a face, the oncologist has a clinic, and the person whose opinion determined the diagnosis and the treatment was never in the room. That has a practical consequence: patients are handed a document written in a technical language nobody translates for them, and are expected to make decisions on the strength of it. A substantial part of what follows exists to make that document readable.

At Acıbadem International the work is organised into five strands.

• Histopathology — the microscopic examination of tissue, which produces the diagnosis, the grade, the margins and the stage.
• Cytopathology — the examination of cells rather than tissue, from smears, fluids and fine needle aspirates.
• Immunohistochemistry and special stains — the tests that identify what a cell actually is when its appearance alone is not enough.
• Molecular pathology — testing the tumour itself for the alterations that decide whether a targeted drug or an immunotherapy can be used.
• Second-opinion slide review — a sub-specialist re-reading of material already reported elsewhere, which is the service international patients most often need and are least often offered.

Where the borders sit. The unit reports the specimen; the clinical unit decides what to do about the result, and the treating sections are written there rather than repeated here. Prostate biopsy technique and prostate cancer treatment belong with urology. Colonoscopy, polypectomy and the surveillance intervals that follow a polyp result belong with gastroenterology. Cervical screening as a pathway — the smear, HPV testing, colposcopy and LEEP — belongs with gynaecology and obstetrics. Breast biopsy, margin re-excision and anti-HER2 treatment belong with breast health. Marrow disease and its treatment belong with haematology, skin lesion management with dermatology, culture and sensitivity testing with infectious diseases, and inherited-risk testing with medical genetics. What this unit owns is the specimen, the report, and what the words in it mean.

Histopathology (surgical pathology): what happens between the theatre and the report

Histopathology — called surgical pathology in much of the world, and the same discipline under either name — is the microscopic examination of tissue that produces the diagnosis, the grade, the margins and the stage. The gap between a biopsy and its result is the part of the process patients find hardest, largely because nobody explains that the time is being spent on something. It is.

Tissue arrives in formalin, which fixes it — stops it degrading and hardens the proteins so they hold their shape. Fixation is not instant and cannot be hurried past a point; a large specimen fixes more slowly than a needle core, which is one of the reasons a mastectomy takes longer to report than a skin punch.

The fixed specimen is then grossed: examined with the naked eye, measured, described, and cut so that the parts that answer the question end up on a slide. This step decides what the microscope will ever be able to see. If the margin is not sampled where the tumour comes closest to it, no amount of microscopy will recover that information. Selected pieces go into cassettes and are processed overnight into paraffin wax, producing the paraffin block — which is the durable form of the specimen and the reason a case can be reviewed, restained or tested years later.

Sections a few micrometres thick are cut from the block and stained, almost always with haematoxylin and eosin — the H&E stain that gives the familiar blue-and-pink appearance. Haematoxylin stains nuclei blue; eosin stains cytoplasm and collagen pink. Most diagnoses are made on H&E alone.

When appearance is not enough, further rounds follow: immunohistochemistry, molecular testing, or a specialist opinion. Each round is a further cycle of cutting, staining and reporting.

Which answers the question of how long results take, without a number that would be dishonest to give. Turnaround is not a queue; it is a sequence of physical steps, and the report is ready when the last necessary one is done. A small biopsy needing only H&E is at the fast end. Bone and other calcified tissue must be decalcified first, which adds days that cannot be compressed. Anything needing immunohistochemistry adds a round; anything needing molecular testing adds another, and sequencing adds more again. A case sent for external expert review adds transit. The honest answer to why is it taking so long is usually that a further test was added because the first answer was not definitive — which is a reason to be reassured rather than alarmed.

Reading your own pathology report

Most modern cancer reports are synoptic: a structured checklist rather than free prose, following an internationally agreed protocol so that no required element is omitted. That is why the document reads like a form. It is a form, deliberately, and its structure is the same worldwide.

A report generally moves through the same sequence.

• Specimen and clinical information — what was received, from where, and the question asked.
• Macroscopic description — what the specimen looked like and measured before slides were made.
• Microscopic description — what was seen under the microscope, in varying detail.
• Ancillary studies — immunohistochemistry, molecular results, or a note that they are pending.
• Diagnosis or conclusion — the operative section. When the body of a report and its conclusion appear to disagree, the conclusion is the report.
• Synoptic summary — type, grade, size, invasion, margins, nodes and stage as discrete fields.

Three habits make the document less frightening. Read the conclusion first, not the microscopic description — the descriptive middle is written for clinicians and contains alarming-sounding words that are simply anatomy. Notice which fields say cannot be assessed, since an unassessable field means the specimen could not answer that question, not that the answer was bad. And treat any result marked pending as genuinely absent rather than implied.

Two conventions confuse people reliably. A prefix of p before a stage means the stage was established on the specimen rather than on imaging, and yp means it was established after chemotherapy or radiotherapy had already been given, which makes it not directly comparable with a pre-treatment stage. And negative in a pathology report is good news: it means the thing being looked for was not found.

Benign, malignant, and the words in between

The distinction patients want is binary and the vocabulary is not, because tissue genuinely occupies intermediate states.

Benign vs malignant: the word that separates them

Benign means the growth does not invade surrounding tissue and does not spread. It may still need removing — for size, symptoms or uncertainty — but it is not cancer. Malignant means the cells invade beyond their normal boundary and are capable of spreading. In a benign vs malignant comparison the single most important word on the report is invasive: everything else is supporting detail.

Dysplasia is the commonest of the intermediate words and the most misread. It describes cells that are abnormal in appearance and organisation but have not invaded. It is graded, usually as low grade dysplasia or high grade dysplasia, and the grade carries most of the meaning: low-grade dysplasia is frequently monitored, while high-grade dysplasia is usually treated because the risk of progression is materially higher. Dysplasia is not cancer. It is also not nothing, and the distinction between those two errors is the reason the word is graded at all.

Carcinoma in situ is the far end of that spectrum: malignant-looking cells that have not crossed the basement membrane. It cannot metastasise while it remains in situ, which is why it is treated with the intention of cure. In the breast this appears as DCIS — ductal carcinoma in situ — and as LCIS, lobular carcinoma in situ, which despite the name behaves as a risk marker rather than as a direct precursor and is managed differently.

Atypia and atypical cells are weaker terms and mean what they say: cells that do not look normal, in a context that does not permit a firmer statement. Atypia appears when there is inflammation, healing or sampling artefact as well as when there is genuine neoplasia, and the usual consequence is a repeat sample or an additional stain rather than treatment.

Two further words appear constantly. A neoplasm is any new growth, benign or malignant — the word carries no verdict. Hyperplasia is an increase in cell number with normal architecture, and is usually a response to something rather than a neoplasm at all.

Tumour grade (tumor grade): differentiation, mitotic rate and Ki-67

Tumour grade (tumor grade) describes how far the cells have departed from the tissue they came from. It is not the stage. Grade is about what the cells look like; stage is about how far the disease has travelled. A small tumour can be high grade and a large one low grade, and treatment decisions frequently turn on the combination rather than on either alone.

Differentiation

The oldest measure is differentiation. A well differentiated tumour still resembles its tissue of origin and generally behaves less aggressively. Moderately differentiated sits between. Poorly differentiated has largely lost that resemblance, and undifferentiated or anaplastic has lost it entirely — which can make the origin itself hard to determine, and is one of the situations that requires immunohistochemistry to resolve.

Mitotic rate and mitotic index

The mitotic rate counts cells caught in the act of dividing, in a defined area. The mitotic index expresses the same idea as a proportion. Both are proxies for how fast a tumour is growing, and both are components of formal grading systems rather than standalone verdicts.

Ki-67

Ki 67 is an immunohistochemical marker present in cells that are actively cycling and absent in resting cells, reported as the percentage of tumour nuclei that stain. It is a proliferation index — a measure of how much of the tumour is dividing. It is genuinely useful in some settings, notably neuroendocrine tumours, where it is part of the formal grade. It is also one of the least reproducible numbers in pathology: the value depends on which field was counted and how, and small differences between reports are frequently method rather than biology. A Ki-67 result is read alongside the rest of the report, never as a score in isolation.

The Nottingham grade in breast cancer

Breast cancer uses the Nottingham grade, also called the Nottingham score or the modified Bloom-Richardson system. It adds three components — how well the tissue forms tubules, how variable the nuclei are, and the mitotic count — each scored from 1 to 3, giving a total from 3 to 9 which becomes grade 1, 2 or 3. Grade 1 is the least aggressive appearance and grade 3 the most. What that grade means for treatment sits with breast health, alongside receptor status and the genomic assays.

The Gleason score and grade groups in prostate cancer

The Gleason score is the most searched and most misunderstood number in pathology, and its structure is the reason.

It does not grade cells; it grades architecture — the pattern the glands form. Patterns are numbered 1 to 5, from glands still recognisably glandular to sheets of cells with no gland formation at all. In practice only patterns 3, 4 and 5 are reported in contemporary prostate pathology.

The score is the sum of two pattern numbers: the most common pattern plus the second most common. This is why the order matters and why gleason 3+4 and gleason 4+3 are different diseases despite both totalling 7. In 3+4, the less aggressive pattern predominates; in 4+3, the more aggressive one does. Reporting them as a single number of 7 discards the more important half of the information, which is precisely the error the system was restructured to prevent.

That restructuring produced the grade group, now reported alongside the score, running from 1 to 5.

• Grade group 1 — Gleason 6 (3+3). The lowest grade reported. Frequently managed by active surveillance rather than immediate treatment.
• Grade group 2 — Gleason 7 (3+4).
• Grade group 3 — Gleason 7 (4+3).
• Grade group 4 — Gleason 8.
• Grade group 5 — Gleason 9 or 10.

Two points about grade group 1 are worth stating plainly, because they cause avoidable distress. A score of 6 sounds like a 6 out of 10 and is not — it is the bottom of the reported range, since patterns 1 and 2 are no longer diagnosed. And the reason the grade group system was introduced was to stop patients reading their lowest-possible grade as a middling one.

Gleason grading is also the setting where second-opinion review changes management most often, because the boundary between pattern 3 and pattern 4 is a judgement about gland architecture on which experienced pathologists genuinely differ — and that particular boundary is the one separating surveillance from surgery. Biopsy technique, imaging and the treatment decision itself belong with urology.

Margins: what clear, close and involved actually mean

When something is cut out, the pathologist inks the outer surface of the specimen before cutting it into slices. That ink marks the true edge — the surface the surgeon cut along — so that under the microscope it is possible to say how near the tumour came to it. Without inking, the question cannot be answered at all, which is why the specimen must reach the laboratory oriented and intact rather than divided.

The report then states one of three things. A negative margin, also called clear or R0, means no tumour reaches the inked surface. A close margin means tumour does not reach the ink but comes near it, with the distance stated in millimetres. A positive or involved margin means tumour cells are present at the inked edge.

Three qualifications matter, and all three are routinely lost in translation.

First, close is not a fixed distance. What counts as an adequate clearance differs by organ and by tumour type — the number that is reassuring in one site is inadequate in another — so a margin measurement is meaningful only against the standard for that specimen, which is why the report states the distance rather than a verdict.

Second, a positive margin does not mean tumour was left behind with certainty. It means tumour was present at the cut surface of what was removed. Whether disease remains depends on what was underneath, and further surgery frequently finds nothing.

Third, some margins cannot be cleared by more surgery because there is nothing further to take — a surface against the chest wall or a fascial plane, for instance. In those cases the finding informs radiotherapy planning rather than a second operation.

R0 resection is the summary term for a complete removal with clear margins microscopically; R1 means microscopic residual disease at a margin, and R2 means visible residual disease. What an involved margin leads to — re-excision, radiotherapy or observation — is a decision for the treating unit, and for breast surgery specifically it is covered by breast health.

Lymphovascular and perineural invasion

Two findings appear on most cancer reports and are rarely explained, despite altering treatment.

Lymphovascular invasion — sometimes written as angiolymphatic invasion, and abbreviated LVI — means tumour cells were seen inside lymphatic channels or small blood vessels within or around the tumour. It matters because those channels are the route by which tumour reaches lymph nodes and beyond. Its presence indicates that the route has been used, not that spread has occurred; many patients with lymphovascular invasion have negative nodes. It commonly influences whether adjuvant treatment is offered.

Perineural invasion — PNI — means tumour cells were seen tracking along the sheath of a nerve. Nerves run in loose planes that offer little resistance, so this is a route by which a tumour extends further than its visible edge suggests. It is particularly significant in prostate, pancreatic, head and neck and certain skin cancers, and it is one reason a tumour can recur at a margin that looked adequate.

Neither finding is a stage and neither is a grade. Both are risk features, reported because they change the estimate of how likely disease is to return.

Two related terms sit nearby. Extranodal extension means tumour within a lymph node has broken through the node’s capsule into surrounding fat — a poorer sign than node involvement alone, and a formal staging element in several cancers. Tumour budding (tumor budding) describes small clusters of tumour cells detaching at the invasive front, most established as a risk factor in colorectal cancer.

pTNM: the pathological stage, and how it differs from the clinical stage

TNM staging describes the anatomical extent of disease in three parts: T for the primary tumour, N for regional lymph nodes, M for distant metastasis. Pathology supplies T and N from the specimen, which is why the report carries a pathologic stage written with a p prefix — pT2 pN0 rather than T2 N0.

The difference between clinical and pathological staging is not academic. A clinical stage is an estimate from examination and imaging before treatment. A pathological stage is a measurement of what was actually removed. They disagree in both directions often enough that the pathological stage supersedes the clinical one wherever it exists.

What determines T varies by organ, and this is a frequent source of confusion: in some sites T is governed by size, in others by depth of invasion through the wall, and in others by which structures are involved. A T-number from one organ carries no meaning in another. Melanoma is the clearest illustration, because there T is set almost entirely by a single measurement made under the microscope — the Breslow thickness, the depth in millimetres from the top of the epidermis to the deepest tumour cell, which outperformed the older Clark level of anatomical layer and replaced it. What that thickness means for surgical margins and sentinel node sampling belongs with dermatology.

N is determined by examining the nodes submitted. Two elements are reported: how many nodes contained tumour and how many were examined in total. The denominator matters — a report of nought out of three nodes is a weaker reassurance than nought out of twenty-two, because the chance of having missed an involved node depends on how many were assessed. Sentinel lymph node sampling is the technique designed to reduce that trade-off: the first node or nodes draining the tumour are identified and removed selectively, then examined in greater depth than would be practical across a full clearance.

M is almost always determined by imaging rather than by pathology, so a report will frequently state that M cannot be assessed from the specimen. That is a statement about the specimen, not about the patient.

Two further conventions. A yp prefix marks a stage established after neoadjuvant chemotherapy or radiotherapy, and it is not comparable with a stage from an untreated specimen. And where a tumour has responded completely to that treatment, the report may describe a pathologic complete response — no residual viable tumour identified in the specimen, which is the strongest single indicator that the treatment worked.

Polyp reports: tubular, villous and serrated

A polyp removed at colonoscopy generates one of the most commonly received and least understood reports in medicine, because the vocabulary sounds ominous and mostly is not.

A tubular adenoma is the commonest type. It is a benign neoplasm — a growth with the potential to progress to cancer over years if left, which is precisely why it was removed. Finding one is the screening programme working, not failing. Villous adenoma describes a frond-like architecture carrying a higher risk of progression, and tubulovillous sits between the two.

A hyperplastic polyp is not an adenoma and, in the common small distal form, carries essentially no progression risk. A sessile serrated adenoma — also called a sessile serrated lesion — looks bland but follows a different molecular route to cancer, which is why it is treated as significant despite its unremarkable appearance.

Three fields on the report carry most of the weight: the type, whether dysplasia is present and at what grade, and whether the polyp was removed completely. A tubular adenoma with low-grade dysplasia is an ordinary finding. High-grade dysplasia in a polyp is still not invasive cancer — the cells have not crossed the muscularis mucosae — but it changes what happens next.

The word that changes the situation entirely is invasive. An adenoma containing invasive adenocarcinoma is a cancer, and is assessed for depth of invasion, grade, lymphovascular invasion and the margin of the polypectomy — the four features that determine whether removal at colonoscopy was sufficient or whether surgery is needed.

How often the next colonoscopy should happen is determined by the number, size and type of polyps found, and those surveillance intervals belong with gastroenterology, which also covers the procedure itself and Barrett’s oesophagus.

Cervical cytology: ASCUS, LSIL, HSIL and CIN

Cervical results use two different systems, one for cells and one for tissue, and the overlap between them is the main reason the result is confusing.

The smear reports cells, using the Bethesda system. ASCUS — atypical squamous cells of undetermined significance — is the deliberately non-committal category: the cells are not normal, and not abnormal enough to classify. It is the commonest abnormal pap smear result, and the majority of cases resolve without treatment, which is why HPV testing is used to sort it rather than immediate colposcopy. LSIL indicates a low-grade squamous intraepithelial lesion, usually reflecting active HPV infection; HSIL indicates a high-grade lesion, which is the result that reliably leads to colposcopy. AGC, atypical glandular cells, is uncommon and taken more seriously than its squamous equivalent.

The biopsy reports tissue, using CIN — cervical intraepithelial neoplasia — graded 1 to 3 by how much of the epithelial thickness is affected. CIN 1 involves the lower third and usually regresses on its own. CIN 2 is intermediate and the least reproducible of the three. CIN 3 is full-thickness and is treated. None of the three is cancer; all describe changes confined above the basement membrane.

Two sources of confusion are worth naming. A smear result and a biopsy result are not the same measurement, so a low-grade smear followed by a higher-grade biopsy is not an error — it is the biopsy answering a question the smear could only guess at. And p16 immunohistochemistry is frequently used to resolve borderline CIN 2, because a p16-positive lesion behaves as high grade regardless of its appearance.

The screening pathway itself — smear intervals, HPV testing, colposcopy and LEEP — belongs with gynaecology and obstetrics.

Cytology and fine needle aspiration

Cytology examines individual cells rather than intact tissue. That is its advantage and its limitation in a single sentence: obtaining cells is quick, cheap and minimally invasive, but without architecture some questions cannot be answered.

Samples come from three routes: exfoliative, where cells shed naturally and are collected — cervical smears, urine, sputum; fluids, where cells are recovered from pleural, peritoneal or cerebrospinal fluid; and fine needle aspiration, where a fine needle draws cells directly from a lump.

What fine needle aspiration can and cannot settle

Fine needle aspiration is excellent at distinguishing a cyst from a solid mass, at confirming a recurrence or a metastasis where the original diagnosis is already known, and at identifying an infection. It is weaker wherever the diagnosis depends on whether cells have invaded, because invasion is a relationship between cells and the tissue around them and an aspirate has separated them from it. The classic instance is the thyroid follicular lesion, where distinguishing an adenoma from a carcinoma requires seeing capsular or vascular invasion — which is why some thyroid nodules proceed to surgery for diagnosis rather than for treatment.

An inconclusive aspirate is therefore a recognised outcome rather than a failure, and the usual answer is a core needle biopsy, which takes a cylinder of tissue with its architecture intact.

The Bethesda categories for thyroid nodules

Thyroid aspirates are reported in six graded categories — non-diagnostic, benign, atypia of undetermined significance, follicular neoplasm, suspicious for malignancy, and malignant. What each category means for the next step — surveillance, repeat sampling or surgery — is the decision owned by the endocrinology unit’s nodule pathway. The system exists to convert an impression into a defined category with an agreed consequence, so that atypical means the same thing in every hospital that uses it.

Cell blocks and liquid-based preparations

A cell block is made by concentrating the cells from an aspirate or a fluid and processing them into paraffin like a tissue specimen. It is a useful step because a cell block can be stained by immunohistochemistry and tested molecularly, which a smear cannot support well. Liquid based cytology collects cells into preservative fluid instead of smearing them onto a slide, which reduces obscuring blood and debris and allows HPV testing from the same sample.

Immunohistochemistry: how a cell is identified

Immunohistochemistry — IHC — uses antibodies to detect specific proteins in a tissue section, producing a visible colour wherever the protein is present. It is the single most important development in diagnostic pathology of the last several decades, because it converts a question about appearance into a question about what a cell is actually making.

It answers three kinds of question.

• What is this? A poorly differentiated tumour that has lost its resemblance to any tissue can still be identified by the proteins it expresses. This is how a metastasis of unknown origin is traced back to a likely primary site.
• Is this benign or malignant? Certain markers separate lesions that look alike — p16 in cervical disease, or markers that distinguish a reactive process from a neoplastic one.
• Will a treatment work? Oestrogen receptor (estrogen receptor) and progesterone receptor status in breast cancer, HER2, and PD-L1 in several tumour types are all measured by immunohistochemistry and determine drug eligibility directly.

Common markers and what they indicate

A handful appear on reports repeatedly. Cytokeratins indicate epithelial origin and therefore carcinoma. CD20 marks B lymphocytes and CD3 marks T lymphocytes, which is how a lymphoid infiltrate is characterised. p53 shows an abnormal pattern where the gene is mutated. p16 acts as a surrogate for high-risk HPV activity. Ki 67 measures proliferation, as covered under grade. S100, SOX10 and Melan-A indicate melanocytic origin. TTF-1 suggests lung or thyroid, and GATA3 suggests breast or urothelium.

Why a marker is never read alone

No marker is entirely specific. Every antibody stains something other than its intended target somewhere, which is why pathologists use panels rather than single stains and why the pattern of positives and negatives carries the meaning. A single positive result quoted in isolation is close to uninterpretable, and this is the commonest way IHC is misread by people outside the specialty.

Interpretation also depends on where the staining sits — nuclear, cytoplasmic or membranous — and on its intensity and extent. HER2 is the clearest example: it is scored 0, 1+, 2+ or 3+ on the completeness and intensity of membrane staining, and a 2+ result is equivocal by definition and must be resolved by in situ hybridization, usually the FISH test, which counts gene copies directly rather than measuring protein.

Immunofluorescence works on the same antibody principle but uses a fluorescent label read under a fluorescence microscope. It remains standard for kidney and skin immune-mediated disease, where the pattern of immune deposits is itself the diagnosis. Alongside these sit the older special stains, which are chemical rather than antibody-based and remain in daily use — among them Congo red, which shows the diagnostic apple-green birefringence of amyloid.

Flow cytometry and immunophenotyping

Flow cytometry analyses cells in suspension rather than on a slide. Cells are labelled with fluorescent antibodies and passed one at a time through a laser, and the instrument records what each individual cell is expressing. Where immunohistochemistry shows a few markers across intact tissue, flow cytometry measures many markers simultaneously on many thousands of individual cells.

That makes it the technique of choice for blood, bone marrow and lymph node suspensions, and it is central to diagnosing and classifying leukaemia (leukemia) and lymphoma. Immunophenotyping is the name for what it produces: the pattern of markers that identifies which cell type a population belongs to, and whether that population is clonal — derived from a single cell, which is the essential feature separating a malignant proliferation from a reactive one.

It has two further routine uses. It detects small abnormal populations at levels well below what a microscope can resolve, which is the basis of measurable residual disease assessment after treatment. And it works on fluids, so a lymphomatous effusion or cerebrospinal fluid can be assessed directly.

Its limitation is the mirror of its strength: it requires a suspension of intact cells, so a sample must be fresh and unfixed. A specimen already in formalin cannot be assessed this way, which is why a decision to send fresh material has to be made at the time of the procedure rather than afterwards.

The related sub-specialty is haematopathology (hematopathology), which integrates the marrow appearance, flow cytometry, cytogenetics and molecular results into one diagnosis. The diseases themselves and their treatment belong with haematology.

Molecular pathology: testing the tumour itself

Molecular pathology tests the tumour’s DNA and RNA for alterations that predict whether a specific drug will work. This is where a pathology report stopped describing a disease and started selecting a treatment. Collectively this work is biomarker testing: measuring a feature of the tumour in order to choose a therapy rather than to name a diagnosis.

Mismatch repair and microsatellite instability

Microsatellite instability — MSI — and mismatch repair deficiency are two ways of measuring the same defect: a failure of the system that corrects DNA copying errors. Mismatch repair status is assessed by immunohistochemistry for four proteins, and loss of any of them indicates deficiency; MSI is assessed by molecular testing. A tumour reported as MSI high or mismatch-repair deficient accumulates large numbers of mutations, which makes it visible to the immune system and frequently responsive to immunotherapy. The same result also raises the possibility of Lynch syndrome, which is inherited — the reason this particular test crosses into inherited-risk territory.

PD-L1 and immunotherapy eligibility

PD-L1 is a protein tumours use to switch off the immune cells attacking them, and its measurement determines eligibility for several immunotherapies. It is reported as a score whose form differs by tumour type — TPS counts tumour cells only, CPS counts tumour and immune cells together — so a PD-L1 number is meaningless without knowing which score was used and which drug it was scored for. Tumour mutational burden (tumor mutational burden) counts mutations per megabase and is used similarly, on the principle that a heavily mutated tumour presents more for the immune system to recognise.

Targetable alterations by tumour type

Lung adenocarcinoma is the most heavily tested, with EGFR mutation, ALK and ROS1 rearrangements, BRAF mutation and several others each corresponding to a specific drug class. Colorectal cancer is tested for KRAS mutation, NRAS and BRAF, where the result determines whether an anti-EGFR antibody can be used at all. Melanoma is tested for BRAF. Gastric and oesophageal (esophageal) cancers are tested for HER2 and PD-L1.

Sequencing panels and liquid biopsy

Next generation sequencing examines many genes at once rather than one at a time, which is now the practical approach when several alterations are relevant. A liquid biopsy detects circulating tumour DNA — ctDNA — in blood instead of tissue. It is genuinely useful where tissue is unobtainable or exhausted, and for tracking resistance as it emerges. Its limitation is that a negative result does not exclude an alteration, because not every tumour sheds detectably; a negative liquid biopsy is not equivalent to a negative tissue test.

Gene expression assays in breast cancer

Oncotype DX and MammaPrint measure the activity of a defined set of genes to estimate recurrence risk and the likely benefit of chemotherapy. They answer a different question from the ones above: not which drug, but whether a drug is worth giving. How the result is used in treatment planning belongs with breast health.

Tumour testing is not genetic testing

This distinction is the one most often lost, and the consequences of losing it run in both directions.

Somatic testing examines the tumour. The alterations it finds arose in that tumour during a person’s life, are present only in it, cannot be inherited and cannot be passed on. Every test described under molecular pathology is of this kind.

Germline testing examines the DNA a person was born with, from blood or saliva. What it finds is present in every cell, may have been inherited, and may be shared with children, siblings and parents.

Two errors follow from confusing them. A patient told their tumour carries a BRCA alteration may conclude their children are at risk when the alteration is somatic and confined to the tumour. And a patient told their tumour testing was normal may conclude that inherited risk has been excluded, when a somatic panel was never designed to answer that question.

Some results legitimately trigger the second test. A mismatch-repair-deficient tumour raises the question of Lynch syndrome; certain tumour BRCA findings raise the question of an inherited variant. In those situations the tumour result is a reason to consider germline testing, not a substitute for it — and germline testing carries implications for a whole family, which is why it is done with counselling rather than as a laboratory add-on. That work belongs with medical genetics. Chromosome-level analysis — karyotyping and cytogenetics — sits between the two, most often used in haematological disease as part of the diagnosis itself.

Frozen section: an answer while the operation is still running

A frozen section is a diagnosis made in minutes while the patient is still anaesthetised. Tissue is frozen solid rather than processed into paraffin, sectioned on a cryostat, stained rapidly and examined immediately, so that the surgeon can act on the answer during the same operation.

It is used for a small number of specific questions: whether a margin is clear before closing, whether a lymph node contains tumour, whether a lesion is malignant at all when the plan differs radically depending on the answer, and whether the tissue removed is even the right tissue — confirming parathyroid rather than fat or node, for instance.

Its limitations are real and are the reason it is not used more widely. Freezing distorts tissue, so fine cytological detail is worse than on a paraffin section. Only a small sample can be examined in the time available. Immunohistochemistry and molecular testing are unavailable in the moment. And freezing consumes tissue that might have been needed for the definitive assessment.

For those reasons a frozen section result is provisional, and every case is processed conventionally afterwards. The final report occasionally differs from the frozen section — a recognised and accepted feature of the technique rather than an error, which is why frozen section is reserved for questions where an immediate answer genuinely changes the operation.

Second-opinion review of slides reported elsewhere

A pathology second opinion is a formal sub-specialist re-examination of material already reported, producing a new written report rather than an informal comment. It is the service international patients most often need, and the one they are least often told exists — largely because the original diagnosis is presented as a fact rather than as an interpretation, which is what it is.

Review changes something a meaningful proportion of the time, and for structural reasons rather than because the first pathologist was careless.

• Sub-specialisation. A general pathologist reports every organ. A sub-specialist reports one, and pattern recognition in diagnostic microscopy is built on volume within a narrow domain. The categories where this matters most are known: lymphoma classification, soft tissue and bone tumours, melanocytic lesions, and the Gleason pattern 3 against pattern 4 boundary.
• Ancillary testing. Additional immunohistochemistry or molecular testing performed on review frequently resolves what the original stains could not, particularly where the first report ended in a descriptive rather than a definitive diagnosis.
• Newer classifications. Diagnostic categories are periodically revised, and an older report may use a category that has since been split, merged or redefined.
• Clinical context. A diagnosis reported without the full clinical picture may read differently when that picture is supplied.

What a review needs is the material, not the paperwork alone. The essential items are the stained slides, or the paraffin blocks, which permit new sections and further testing and are the more useful of the two. Photographs of slides are not reviewable. The original report should accompany them, with any imaging and the clinical question. Digital pathology changes the logistics where it is available: whole slide imaging scans a glass slide into a high-resolution digital file, and telepathology allows that file to be reviewed remotely, so that in scanned cases the data travels rather than the specimen.

Two honest limits. A review cannot recover what the specimen never contained — if the biopsy missed the lesion, the answer will be that a further sample is needed rather than a different diagnosis. And a review that confirms the original report is not a wasted exercise: independent confirmation is frequently what allows someone to proceed with treatment instead of continuing to look for a different answer.

How a laboratory checks its own work

A diagnosis nobody can audit is a diagnosis nobody should rely on, so a substantial part of laboratory work is checking the laboratory.

Identification comes first and is the failure with the worst consequences: a specimen attributed to the wrong patient produces a perfect report about someone else. Specimens are tracked by identifiers from the moment of collection through every transfer, and the discipline around that is as important as anything done with a microscope.

Stains are controlled against known material. Every immunohistochemical run includes a control section known to contain the target protein, so that a negative result can be distinguished from a failed stain — without it, absence of colour is uninterpretable.

Difficult cases are reviewed internally before they are issued. Categories with recognised interobserver variation — lymphoma, soft tissue and melanocytic lesions among them — are seen by more than one pathologist, and cases outside a reporter’s sub-specialty are referred rather than reported at the limit of familiarity.

Reports are correlated against everything else known about the case. A pathological result that contradicts the imaging or the clinical picture is investigated rather than issued, because such a discrepancy is one of the more reliable indicators that something has gone wrong.

Cancer reporting follows internationally published protocols, which is why synoptic reports contain the same required fields wherever they are produced — and which is what makes a report from one country usable by a team in another. External quality assessment adds the outward-facing half: laboratories are sent material of known diagnosis and their answers are compared against other laboratories, so that performance is measured against the field rather than against itself.

What pathology cannot do

It cannot diagnose what was not sampled. A biopsy examines the tissue it contains, and a benign result on a needle core that missed the lesion is a true statement about the wrong tissue — which is why a result inconsistent with the imaging leads to resampling rather than reassurance.

It cannot always give a single answer. Some entities sit genuinely on a boundary, and a report that says suggestive of or cannot be excluded is being accurate. Forcing false certainty onto an ambiguous appearance causes more harm than admitting the ambiguity.

It cannot predict an individual’s course. Grade, stage and biomarkers describe populations and shift probabilities. They do not tell one person what will happen to them, and a report should not be read as a forecast.

It cannot recover tissue that has been used. Material is finite, and every stain and every molecular test consumes some. This is a real constraint on small biopsies, and it is the reason testing is sequenced deliberately rather than requested all at once.

It cannot assess what fixation destroyed. Flow cytometry and some molecular assays need fresh or specially preserved material, and once a specimen is in formalin those options are gone — a decision made in the operating theatre, not in the laboratory.

It cannot replace a genetic assessment. Tumour testing describes the tumour, and the distinction from inherited testing is one this unit states explicitly rather than leaves to be inferred.

Your multidisciplinary team

The pathologist examines the specimen, integrates the ancillary results and issues the report, sub-specialised by organ system — because the difference between a general and a sub-specialist report is frequently the diagnosis itself. The cytopathologist reports smears, fluids and aspirates, and where rapid on-site evaluation is used, confirms during a procedure that the sample obtained is adequate before the needle is withdrawn. The molecular pathologist runs and interprets the tumour genomic testing and takes those results to the tumour board. The biomedical scientist and laboratory technologist prepare the specimen — grossing assistance, processing, sectioning and staining — and the quality of a slide sets the ceiling on what can be read from it. The laboratory quality manager owns identification, controls and external assessment.

Around them, and this is the specialty that touches every other: medical oncology and the tumour boards where the report drives treatment, general surgery and every surgical unit that submits specimens, radiology for image-guided sampling and for the correlation that validates both, haematology for marrow and lymphoid disease, breast health, urology, gastroenterology, gynaecology and dermatology for the specimens they generate in the greatest numbers, medical genetics where a tumour result raises an inherited question, and infectious diseases where the question is microbiological rather than neoplastic.

The international patient journey

This unit travels the way radiology does: in most cases the material travels and the patient does not.

The largest pattern by far is the remote slide review. It requires the glass slides or, better, the paraffin blocks, together with the original report and the clinical question — and it concludes with a written report rather than a journey. Where a case has been digitally scanned, even the material stays where it is.

The second pattern is a diagnosis that was never completed. A report ending in a descriptive line rather than a definitive diagnosis, or one issued before the immunohistochemistry or molecular testing that the case needed, can frequently be finished from the existing blocks without a new procedure.

The third is testing that was not available where the patient was treated — a molecular panel, a specific assay, or sub-specialty expertise in a category where general reporting is known to be unreliable. Here too the block is what matters.

The fourth, and the only one that requires travel, is a new diagnosis: a lesion that has not been sampled, or one where the previous sample was inadequate. That is planned with the unit performing the procedure rather than with the laboratory.

Three practical notes. Blocks are more useful than slides, because new sections and further testing can be cut from a block and cannot be cut from a slide that already exists. Slides and blocks are usually retained by the originating laboratory and released on request to the patient or to a receiving laboratory, so obtaining them is an administrative step worth starting early rather than a barrier. And the original report should travel with the material, because a review with no knowledge of what was previously concluded, or of what question was being asked, is a weaker exercise than one that has both.