Mediastinoscopy & Lymph Node Biopsy
When Needle Sampling Has Not Settled It

Endobronchial ultrasound with transbronchial needle aspiration is the first-line test for sampling lymph nodes in the centre of the chest, and this page is not an argument against it. A camera with an ultrasound probe on its tip is passed into the airway, the nodes are seen through the airway wall, and a needle is passed into them under direct ultrasound vision. It reaches the paratracheal and subcarinal stations reliably, it is performed under sedation without an incision, and its diagnostic yield for common pathology is high. For staging lung cancer and confirming metastatic disease in the glands, it answers the question in most patients. It has correctly displaced mediastinoscopy as the initial procedure.

Where it stops is more specific than most patients are told:

A mediastinoscopy answers each of those differently, for one physical reason: it removes whole lymph nodes rather than sampling them through a needle. The difference is not that one produces a better specimen in general. It is that mediastinoscopy provides an intact node, including its architecture, when that is what the diagnosis depends on.

EBUS is not static, and a page that ignored this would be out of date within a year. EBUS-guided transbronchial mediastinal cryobiopsy passes a fine cryoprobe into the node through the airway, freezes a core of tissue and retrieves it. The specimen is substantially larger than a needle aspirate and considerably better preserved, which addresses two of the four limitations set out above.

The published results are strong, particularly in exactly the disease where surgical biopsy has traditionally been argued for. A prospective multicentre study in suspected lymphoproliferative disorders found a diagnostic yield of 80% for cryobiopsy against 52% for needle aspiration, rising to 85% against 41% in confirmed cases, with 80% of non-diagnostic aspirates subsequently diagnostic on cryobiopsy. A systematic review reported cryobiopsy diagnostic in 87% of lymphomas against 12% for needle aspiration, and able to characterise every lymphoma subtype. A randomised trial in non-metastatic lymphadenopathy found 97.1% against 79.9%.

A modified Delphi consensus statement published in CHEST in 2026, involving 32 physicians across four survey rounds, provides the first formal expert guidance for the use of transbronchial cryobiopsy outside interstitial lung disease, including convex-probe EBUS-guided cryobiopsy and specimen handling. It exists because practice has spread faster than standardisation, and it reached consensus on some questions and explicitly failed to reach it on others.

The honest position is that cryobiopsy narrows one of the traditional reasons for proceeding to mediastinoscopy. It does not close the indications where definitive surgical staging is what the treatment decision requires, where an intact node is specifically needed, or where the endoscopic route has already been attempted without an answer. It is also concentrated in centres with established EBUS and cryobiopsy expertise, so the alternative available to a particular patient depends on where they are being investigated.

Five

Selected diagnostic problems where a whole node is specifically required. Not everything falls into four neat categories. Suspected recurrence in a previously treated field, a node that has been sampled twice without a clear answer, an unusual differential where the pathologist has asked for more material, disease after previous chemotherapy or radiotherapy where the tissue is necrotic and a needle keeps returning debris. In each case the question is the same: what will the laboratory need in order to give a definite answer, and what is the least invasive way of providing it?

Lymphoma is one of the situations in which preserving tissue architecture can be decisive. The diagnosis frequently depends not only on the abnormal cells but on how they are arranged within the node — whether the pattern is follicular or diffuse, how far the normal structure has been effaced, the relationship of the infiltrate to the capsule and the sinuses. Subtyping under the current WHO classification then requires immunophenotyping, flow cytometry and molecular studies, each of which consumes material.

The practical consequence is that the biopsy method should be chosen in discussion with the haematologist and the haematopathologist rather than assumed. Some subtypes are reliably diagnosed on small samples; others are not. An inadequate specimen does not simply delay the diagnosis — it risks an incomplete or incorrect one, and treatment in lymphoma is subtype-specific.

Excisional biopsy remains an important approach where an intact node is required, and where the abnormal nodes are confined to the mediastinum a cervical mediastinoscopy is the route that reaches them. At the same time, EBUS-guided cryobiopsy is increasingly able to provide larger and more architectural samples from selected mediastinal nodes, and where it is available it is a reasonable step before surgery.

What should not happen is a succession of small-volume samples taken in sequence, each inadequate for the same reason, while the diagnosis waits. If the first sample was reported as insufficient for lymphoma assessment, the next attempt should be chosen to answer that specific problem.

A diagnosis of adenocarcinoma is no longer the end of the tissue requirement. The same specimen may need to support immunohistochemistry to confirm the subtype, PD-L1 scoring, and broad molecular profiling across a panel that now routinely includes EGFR, ALK, ROS1, BRAF, KRAS, MET, RET and NTRK, with next-generation sequencing requiring a minimum tumour cell content and DNA yield. Occasionally the tumour volume recovered will not stretch to all of it.

EBUS-TBNA supports molecular testing in most patients, and it is important to say so plainly rather than to imply that needle sampling is inherently inadequate for modern oncology. The situation this section describes is narrower: the laboratory has reported insufficient tumour or insufficient material, and the treatment decision is waiting on a result that cannot be produced from what was sent.

In that position, obtaining a larger specimen may be necessary rather than repeating the same small-volume sampling strategy and hoping for a better yield. Dr Okiror is asked to perform mediastinoscopy for precisely this reason by oncology colleagues — not because the first test was done badly, but because what is now being asked of the tissue has grown.

The wider sequence of investigations before treatment, and where nodal sampling sits within it, is set out on the lung cancer tests and staging page.

The operation is performed under general anaesthetic with the neck slightly extended. A transverse incision of about 3 cm is made at the base of the neck, sitting in a natural skin crease just above the sternal notch. The strap muscles are separated in the midline and the plane immediately in front of the trachea is opened, first with a finger and then under direct vision, so that the mediastinoscope passes down alongside the windpipe into the centre of the chest.

From that position the nodal stations that matter for staging and for diagnosis are within reach: 2R and 2L in the upper paratracheal region, 4R and 4L lower paratracheal, and station 7 beneath the carina where the trachea divides. Nodes are dissected free and removed whole. Where a lesion is being sampled for diagnosis rather than staged, tissue is taken generously, because the commonest reason for a repeat procedure is a specimen that was too small for what the laboratory was later asked to do with it.

Selected specimens may undergo intraoperative assessment where an immediate result would alter the procedure being performed. That answers a specific question on the day; it is not the final diagnosis. Definitive histopathology, and any required immunohistochemistry, flow cytometry or molecular testing, follow afterwards.

The wound is closed in layers with an absorbable suture placed beneath the skin, so there are no stitches to remove. A drain is not usually required.

Dr Okiror performs mediastinoscopy as a day case. You are admitted in the morning, having eaten nothing from midnight, and seen by the surgeon and the anaesthetist before going to theatre. The operation itself takes under an hour in most cases. You wake in recovery, spend a few hours on the ward, eat and drink, and go home the same day.

You will need a responsible adult to collect you and stay with you overnight, as after any general anaesthetic, and you should not drive, operate machinery or sign anything legally binding for 24 hours. Most people take a few days off work. Discomfort is usually a sore throat from the breathing tube and a sore neck from the dissection, controlled with simple analgesia.

Who stays overnight instead. Patients with significant cardiac or respiratory comorbidity, patients on anticoagulation that requires monitored reversal and restart, those who live a long way from the hospital or have nobody at home, and the occasional patient whose operation proved more difficult than the imaging suggested. That possibility is discussed before the day rather than raised on it.

Day-case status matters more here than it sounds. This is a diagnostic operation performed on someone who does not yet know what they have, frequently while a treatment decision waits. A procedure that costs one day rather than an inpatient admission is easier to accept and quicker to arrange.

Mediastinoscopy is a well-established operation with a low complication rate in experienced hands. It is still an operation rather than a test, and it is consented for as one.

Common and self-limiting: sore throat, a sore or stiff neck for a few days, bruising around the incision, and temporary discomfort on swallowing.

Uncommon: hoarseness, where the nerve supplying the voice box is stretched or irritated during dissection — more relevant on the left, where that nerve loops beneath the aortic arch close to the operative field. It is usually temporary. Wound infection is unusual in a clean neck incision.

Rare but serious: bleeding from the large vessels lying immediately alongside the dissection, injury to the trachea or oesophagus, and pneumothorax. Any of these would convert a short day-case operation into a larger one, potentially requiring sternotomy. They are rare, and they are the reason this operation belongs with surgeons who perform it regularly rather than occasionally.

There is also a diagnostic risk worth naming: a mediastinoscopy samples the stations it can reach. A negative result is strong evidence about those stations and says nothing about nodes elsewhere in the chest. Where the imaging points to a compartment outside its reach, a different route is chosen from the start.

Definitive histopathology usually takes several days. Where immunohistochemistry, flow cytometry or molecular profiling are required, the complete result takes longer — commonly one to three weeks depending on the panel requested. That wait is uncomfortable and it is better anticipated than discovered. The result is discussed at the multidisciplinary team meeting and then gone through with you directly.

Where the answer changes the treatment plan, the referring oncologist, haematologist or physician receives it at the same time, so that the next step is arranged without a further round of appointments.

How the diagnostic routes fit together

Three routes, three different targets. ION navigational bronchoscopy obtains tissue from a lesion out in the lung itself, without an incision. EBUS samples and stages the lymph nodes in the centre of the chest through the airway wall. A mediastinoscopy provides surgical nodal tissue when the endoscopic route has not answered the question. They are complementary rather than competing, and the right one depends on where the tissue has to come from and what the laboratory will be asked to do with it.