Prostate cancer, with a prevalence of 9.6% and 679 000 cases diagnosed in 2002, is the third most common cancer worldwide.1 In the UK, it is the most common male cancer, representing 23% of all new cancer diagnoses, and is a major source of morbidity and mortality, causing 10 000 deaths in 2005.2 Over the past two decades, the incidence of prostate cancer has risen due to its more frequent diagnosis by a combination of serum prostate-specific antigen (PSA) testing, digital rectal examination (DRE), and transrectal ultrasound (TRUS)-guided biopsy. In the UK, 5-year survival has improved from 31% in the 1970s to 71% in 2001,2 a finding attributed to the earlier detection of tumours, many of which behave indolently, and improvements in treatment. Although treatment of such “screening detected” cancers remains controversial, advances in surgical technique, facilitating postoperative preservation of continence and potency, mean that radical prostatectomy (RP) is increasingly offered as treatment of choice. The consequent increase in RP and pelvic lymphadenectomy (PLA) specimens received by pathology departments poses significant resource implications, highlighting the importance of effective and efficient specimen processing.

The biological behaviour of prostate cancer is heterogeneous and its underlying molecular mechanisms are poorly understood. Factors that might reliably predict outcome for individual patients (facilitating management and prognostication) have been identified and evaluated by two recent international consensus conferences.3 4 The first, organised by the College of American Pathologists (CAP),3 analysed and stratified existing predictive factors into one of three categories based on the strength of published evidence and opinions of its Prostate Working Group (table 1). The second, organised by the World Health Organization (WHO),4 concentrated on biopsy derived predictive factors. It ratified the CAP ranking and recommended the mandatory adoption in clinical practice of all category I factors, advised local discretion with regard to category II factors and did not advocate routine use of category III factors. Since many of these factors are derived from the histological examination of prostate specimens, optimal processing remains paramount.

Standardisation of processing and reporting ensures uniformity and comprehensiveness of examination, and is facilitated by using protocols and checklists for macroscopic and microscopic evaluation; this practice is endorsed by the Royal College of Pathologists (RCPath)5 in the UK, the CAP,6 and the Commission on Cancer of the American College of Surgeons.7 In its Minimum dataset for prostate cancer histopathology reports,5 the RCPath outlines evidence-based processing guidelines and defines the minimum standard of information required to report prostate cancer specimens; recognising that the means by which such standards are achieved will be dependent on local facilities, emphasising the need for protocols to be designed and agreed at a local level. Regardless of the protocol used, meticulous examination of RP and PLA specimens remains critical in providing essential information from which to predict patient outcome and determine the need for adjuvant therapy. In this review, we analyse the current literature and define best practice guidelines for the macroscopic handling of RP and PLA specimens.

ANATOMICAL CONSIDERATIONS

The prostate varies widely in size and shape and can be difficult to orientate. Generally, it resembles an inverted cone, the concave base lying in continuity with the bladder neck, and the apex, usually convex, at the external sphincter in continuity with the urogenital diaphragm (fig 1). In the surgical anatomical position, the seminal vesicles lie inferoposteriorly and may be partly embedded within the gland. The urethra passes centrally through the prostate, curving anteriorly at the level of the verumontanum.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Prostate in surgical anatomical position.

Margins

Parts of the prostate designated with margin status include the bladder neck (proximal or basal), apex (distal), tips of the vasa deferentia, and the anterior, lateral and posterior surfaces of the remainder of the gland (circumferential).

Prostate “capsule”

The posterior surface of the prostate is generally flat and smooth because of to the presence of a discrete tissue layer, Denonvilliers’ fascia, separating it from the rectum. Elsewhere, due to the inconsistency of periprostatic fibromuscular tissue, the prostate lacks a definitive histological boundary; thus the term “extraprostatic extension” (EPE) is preferred to “capsular invasion”.

HANDLING RP SPECIMENS

The excised gland may be delivered to the pathology laboratory fresh or in a fixative such as 10% neutral buffered formalin, the volume of which should be at least 300 ml or five to ten times the volume of the specimen. The specimen pot should be adequately labelled and accompanied by a request form providing important background information (box 1). Specimen and patient details should be verified.

Initial macroscopic evaluation

Initial assessment can be performed either on the fixed or fresh specimen. Dried, fixed prostates are more convenient to ink, have fewer surface cracks for ink to penetrate (reducing the risk of false margin positivity), and are easier to slice, producing thinner sections at consistently even levels. Neoplastic tissue is often better appreciated in fixed specimens; although the sensitivity and specificity of such examination is poor.8 Palpation of tumour nodules, particularly ill-defined peripheral zone tumours, is easier in fresh specimens, facilitating harvest of such tissue for research and special studies. Specimens should be handled with care as even minimal disruption of the surrounding scant soft tissue can produce false-positive margins.

Specimen description

The report on the macroscopic features should include a number of key points (box 2). Prostates are irregular structures, symmetrical in only one axis, and linear measurements are often less accurate and reproducible than specimen weight.

Fixation

The simplest method involves immersing the uncut gland in an adequate volume of fixative for 24–48 h. Adequacy of fixation can be limited by the “crust effect”: as fixative penetrates the superficial prostatic tissue, chemical cross-linking of proteins hinders its further diffusion towards the centre, resulting in central autolysis and producing poor microscopic detail and unreliable results with immunohistochemistry. To achieve faster, more uniform fixation the prostate can be injected with up to 100 ml formalin at multiple sites 5 mm apart,9^–15 or subjected to microwave-stimulated fixation.14

Consideration should be given to the routine storage of prostate cancer tissue for subsequent DNA, RNA or proteomic analysis, either as unfixed frozen tissue or fixed in an appropriate medium such as 70% ethanol, as it may find future use in the development of targeted and individualised therapy. Microwave-stimulated formalin fixation allows DNA extraction from paraffin-embedded material, expanding its potential for molecular analysis.14

Recommendations

• Immersion in fixative is adequate for routine histological examination.

• Preservation of fresh frozen or appropriately fixed tissue for future DNA, RNA and proteomic analysis should be considered.

Inking

Given the absence of a discrete prostatic “capsule”, most pathologists use ink to identify margins,14 allowing differentiation of a truly positive margin from an artefact of sectioning. Use of multiple colours to designate laterality (important in the substaging of pT2 tumours) or to distinguish individual margins facilitates specimen orientation in microscopic sections. If partial sampling protocols are employed, discriminant inking of margins can help maintain the anatomical integrity of residual tissue should more tissue be required for examination. Some pathologists use a single colour, particularly if using routine rather than whole-mount sections, and indicate laterality by means of a scalpel nick, taking care to avoid disruption of the margin. The proximity of tumour to a margin is unpredictable on the basis of gross examination, thus the entire gland surface should be inked.9 15 Immersion in a mordant, such as Bouin’s solution or ascetic acid, can help fix ink to the surface.

Recommendation

• The entire surface of the prostate should be inked using at least two colours to indicate laterality on microscopic sections.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Prostate inked with two colours to designate laterality.

Sectioning the prostate

Margins

A positive surgical margin, defined as tumour extending to a surface transected by the surgeon,10 16 17 implies residual tumour within the patient, and confers an increased risk of local tumour recurrence and biochemical (PSA) progression,10 16 and a reduction in disease-specific survival.18 It may also prompt institution of adjuvant therapy.16 19 A common iatrogenic cause of margin positivity is the operative transection of an otherwise gland-confined tumour. This so-called “capsular incision” has been shown to have little effect on prognosis.17 In an area of EPE, however, a positive margin is an adverse prognostic feature,20 shown on multivariate analysis to be an independent predictor of disease progression with rates of up to 60% at 5 years.18 21

Better patient selection for surgery, improvements in operative technique and overall stage migration have reduced the incidence of margin positivity.2 Its presence should still be carefully sought, however, given its prognostic significance and the potential for adjuvant therapy. Adverse tumour characteristics, including large volume, high grade and EPE, increase the probability of finding positive margins.16 Methods of specimen processing also influence this probability, with RPs sectioned at 4–6 mm intervals missing an additional 12% of positive margins compared with those sectioned every 2–3 mm.10 17 18

The use of intraoperative frozen section analysis to assess margin status is controversial. Some find it useful in deciding whether to preserve one or both neurovascular bundles, however, the potential for false positivity and false negativity probably preclude its useful application.17 22^–24

Recommendation

• Intraoperative frozen section analysis of margins is not recommended.

Margin processing

Distal (apical) margin

No precise anatomical definition exists as to what constitutes the apical margin of the prostate and there is little uniformity in its pathological processing, although most agree it should be entirely embedded and examined either as shave (en face) or perpendicular sections.

Shave section

Shave sections generate high rates of margin positivity and are not advocated.10 They are used by some practitioners who accept this high rate of positivity as, although it confers an increased risk of biochemical relapse,25 evidence suggests that a positive distal margin alone confers less of a negative prognostic impact than a solely positive margin at another site.3

A thin (⩽1 mm) tangential shave section is removed from the apex, embedded en face, and designated as positive if neoplastic glands are present anywhere within the section. The theoretical advantage of this method is the availability of the entire distal margin for histological assessment. A disadvantage is its tendency to produce false-positive margins, when tumour close to but not at the margin is present in microscopic sections. The rate of false positivity is dependent on the specific en face method employed and the examining pathologist’s criteria for margin positivity. Few papers comment on whether such en face sections are embedded “cut side up” or “cut side down” and only one26 states whether the first or last section through the block is examined histologically. In the latter paper, Humphrey et al26 ink the distal surface of the shave section and ask the histotechnologist to cut only superficial sections from this surface, ensuring that only the true distal margin is examined microscopically. In his 2006 paper, Srigley refers to excessive block trimming as a potential cause of false positivity,7 implying that only the most superficial or distal section of an en face shave margin is examined histologically.

Perpendicular sections (analogous to cervical cone biopsy processing)

Depending on the study, the distal 4–15 mm of the prostate is transected and sliced parasagittally at 2–5 mm intervals, perpendicular to the inked margin and parallel to the urethra (fig 3). Slices of even thickness (not radially cut wedges) are embedded on edge. The margin, assessed longitudinally, is designated positive if transected neoplastic glands are present at the inked surface. Tumour close to (within 0.1 mm) but not at the margin is not designated positive since it confers a similar prognosis to a negative margin.10 Perpendicular sectioning theoretically reduces the risk of false margin positivity and is the preferred method of processing; however, as only one piece of tissue (5 μm thick) for every 2–5 mm of distal margin is available for histological examination, there is the potential for false negativity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Apical margin processing. The distal 5 mm of the prostate is transected (A) and sliced perpendicularly at 2–5 mm intervals (B).

A variant of perpendicular sectioning involves a thin shave of the apex being radially sliced as a “pie”. Each wedge-shaped fragment is embedded on edge27 and the margin designated as positive if transected neoplastic glands are present at the inked margin. Correct orientation of these small fragments at embedding can be difficult; hence this method is not advocated.

Distal urethral limit

The “distal urethral limit” is not considered equivalent to the distal (apical) margin7 17 and its separate processing, commonly as a core from the distal section, is controversial. Opponents believe urothelium to be unimportant in this context since prostate cancer does not exhibit pagetoid or submucosal spread without involving adjacent tissues.28 Examination of tissue surrounding the urethra (ie, the distal margin) is felt to be more important than examining the urethra itself.10 With fixation, urothelium often retracts into the apical segment, and digging into the prostate to obtain such tissue could disrupt and obscure a truly positive distal margin, or increase the probability of false positivity, and is discouraged.10

Proximal (basal or bladder neck) margin

The proximal margin, comprising thick bundles of smooth muscle in the absence of benign prostatic glands, should be embedded and histologically examined in its entirety using either shave or perpendicular sections as described above. Given the high rate of false positivity with shave sections, perpendicular sections are preferred (fig 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4 Basal margin processing. The proximal 5 mm of the prostate is transected (A) and sliced perpendicularly at 2–5 mm intervals (B).

Vasal margin

The vasa are identified as short stumps medial to the seminal vesicles bilaterally. While they rarely represent an isolated positive margin,10 17 their involvement by tumour correlates with a poor prognosis. Submission of their proximal portions, as shaves or en face sections, is recommended rather than mandatory.7

Recommendations

• Proximal and distal margins should be processed as perpendicular sections, submitting the entire bladder neck and apical portions of the gland for histological examination.

• Use of shave sections for margin assessment is not recommended.

• Separate examination of the “distal urethral limit” is not recommended.

• Vasal margins can be sampled as transverse sections (not mandatory).

Seminal vesicles

Seminal vesicle (SV) invasion (SVI), defined as tumour infiltrating the muscular wall10 17 29 and designated stage pT3b, is uncommon in the absence of EPE and, although it diminishes the likelihood of cure, is not an independent predictor of prognosis.17 Some pathologists regard tumour involving “intraprostatic” portions of the SVs as stage pT3b. As the prognosis for such patients is equivalent to those without SVI, this practice is discouraged.17

Four routes of SVI have been proposed (box 3). The most common (direct extension into seminal vesicle muscle coat from stroma at base of prostate and tumour spread out of prostate into extraprostatic soft tissue and back into seminal vesicle) necessitate tumour invasion via the base of the SV where it joins the prostate. A section through this area represents the minimum amount of tissue required for histological assessment; however, given the adverse prognostic implications of SVI, best practice should include submission of the SVs in their entirety eg by bisecting them longitudinally and submitting both halves in one or two cassettes.7 11 15

Recommendation

• Seminal vesicles should be embedded in their entirety.

The remainder of the gland

The remainder of the specimen is serially sectioned from apex to base, in a plane perpendicular to the rectal surface and parallel to the initial en face section of the proximal margin, using a long, sharp disposable knife, a specifically designed prostate cutting machine, an adapted meat slicer,30 31 or another motor driven cutting device.11 12 15 32 Slice thickness depends on whether whole mount “macroblocks” or routine sections are preferred.

Depending on gland size, whole mount sections allow each complete en face slice to be embedded in a single “macroblock” generating approximately 5–9 slides12 33 with a slice thickness of 4–6 mm. Routine sections, preferred by the majority of pathologists,10 are 2–3 mm in thickness, require each en face slice to be halved or quartered before embedding in standard cassettes, and generate approximately 26 slides.26 Whole mounts are aesthetically pleasing, particularly for digital imaging (eg, for reporting, teaching or publication). They facilitate slice orientation and examination, tumour identification and quantification, assessment of tumour multifocality and margin status and correlation with preoperative imaging; however, all this information can be obtained from routine sections processed in standard sized blocks. The benefits of whole mount sections are countered by higher costs for labour and materials (eg, cassettes, slides and cover slips of larger size), the need for a higher level of histotechnologist skill (ensuring relevant margins and extraprostatic soft tissue are preserved in all sections), and the logistical problem of storing blocks and slides of unconventional size. Thicker sections require more time for fixation and miss 7–15% of EPE and 12% of positive margins when compared with thinner, routine sections.10

Tumour identification

Once sectioned, slices are laid out and labelled sequentially. Diagrams, digital photographs or photocopies can be used to illustrate specimen anatomy, orientation and gross appearance, record the number and location of blocks taken, facilitate identification of multifocality and calculate tumour volume. Many pathologists visually inspect and palpate slices for tumour to guide sampling. Grossly apparent tumour foci are usually greater than 5 mm in dimension and may show asymmetry, colour differences, granularity, and greater solidity compared with adjacent pale, spongy or cystic non-neoplastic tissue.28 34 35 The recent rise in small, asymptomatic, PSA-detected cancers has rendered gross identification of tumour unusual. Such tumours rarely form discrete masses, are often multifocal, and can be obscured by, or indistinguishable from, benign nodular hyperplasia,36 inflammation and atrophy.

Sampling

Any sampling method should allow each tumour to be histologically identified, graded, quantified and staged, and permit assessment of all prognostically important pathological features. Currently, no consensus exists as to whether the whole gland, or a selected fraction of it, should be embedded for these purposes. The RCPath minimum dataset states that where whole mounts are available, small glands should be blocked in their entirety; otherwise representative blocks of the circumferential margin should be taken.5 The rationale for embedding RP specimens in their entirety is the poor reliability of gross examination to accurately identify tumours, particularly those that are small or multifocal, and its purpose is to maximise the amount of tissue available for examination. However, “complete embedding” does not necessarily equate to “complete histological examination”. In a routinely sectioned, completely embedded prostate (yielding about 26 blocks26), examining a single section (5 μm thick )from each 3 mm tissue block would result in the histological review of only 0.17% of embedded tissue.26 28 37 38 Serial or close step sectioning increases the amount of tissue available for examination, significantly enhancing the detection of adverse prognostic features (eg, EPE, margin positivity and SVI) and improves disease-free survival by pathological stage when compared with specific partial sampling techniques.30 Theoretically, however, such sectioning could produce 15 600 slides per case,26 28 which would be impractical and costly.

In an attempt to reduce time and cost, some pathologists utilise a form of partial sampling.14 Given the risk of missing important information, sampling should not be purely mechanical, protocol driven or random, but influenced by relevant preoperative clinical parameters, biopsy histology and, to a limited extent, the gross appearances of the specimen. A variety of sampling schemes have been documented.5^–7 10 17 26 28 30 32^–35 37 39^–43 Most include submission of the entire proximal and distal margins (described above), all or part of the seminal vesicles, and sometimes the tips of the vasa. The remainder of the gland is sampled according to the presence or absence of grossly visible tumour, bearing in mind that only 26% of stage T1c tumours are grossly evident on cut section.10 40 Processing all visible tumour together with margins detects 96% of positive margins and 91% of EPE when compared with complete embedding.10 17 In the absence of macroscopically visible tumour, many schemes involve preferential sampling of the posterior prostate, since most tumours of stage T1c and above arise posteriorly. This will detect 92% of tumours with a Gleason score greater than 7, 86% of those with positive margins, and 84% with EPE.10 17 Since 15–25% of T1c tumours are located in the transition zone anteriorly,10 17 40 additionally sampling at least one anterior block from each side (plus the entire ipsilateral side should either anterior section prove positive) will increase the detection of tumours with Gleason scores greater than 7–98%, and identify 100% of positive margins and 96% of EPE.10 17 Where local constraints preclude complete embedding, use of the partial sampling scheme outlined in box 4 is recommended.

Recommendations

• The prostate should be serially sectioned at 3–5 mm intervals.

• The gland should ideally be submitted in total for comprehensive staging and assessment of margin status.

• Where complete embedding is precluded by local constraints, the partial sampling scheme outlined in box 4 is recommended.

• The use of whole mount or routine sections remains subject to local discretion.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5 Slicing the remainder of the prostate. (A) The gland is serially sectioned at 5 mm intervals using a plastic gauge. (B) Slices are laid out sequentially and cut surfaces examined. (C) Sections placed in blocks for processing.

Further processing

Some pathologists “post-fix” the sliced specimen in formalin for a further 24 h.9 12 15 Sections are dehydrated in graded alcohols, cleared in xylene, and embedded, as flat as possible, in paraffin wax. H&E-stained sections (5 μm thick) are produced for microscopic evaluation. Histotechnologists should be instructed to trim sections such that all inked margins are available for histological examination.

Indications for additional levels or further tissue submission

Practical indications for examining further levels or additional tissue include confirmation of margin status or EPE, and identification of carcinoma if none is found following a positive biopsy (not for RPs following a positive TURP, as tumour may have been completely excised by the procedure). If carcinoma is not identified, improper patient labelling should be considered, the biopsies re-examined and RP tissue blocks flipped and re-cut. If partially sampled, remaining tissue should be examined. Immunohistochemistry (eg, for cytokeratins) can be used to identify “minimal residual cancer”. If carcinoma remains undetected, it is likely to be small and of no clinical significance.12

Harvesting fresh prostate cancer tissue for research

Recent advances have permitted the application of many molecular biological techniques to formalin-fixed, paraffin-embedded tissue; however, fresh tissue remains the ideal. Obtaining a representative tumour sample can be problematic, as prostate cancers can be difficult to appreciate grossly in fresh specimens10 36 39 and are often multiple or multifocal, with grade varying from one focus to another.36 When the fresh prostate is incised, centrally located tissue bulges beyond the plane of section and extraprostatic fibromuscular tissue retracts; a feature commonly exploited in the procurement of fresh tissue.9 10 17 The inked RP is bisected, a shave of protuberant tumour tissue taken, and the gland sutured27 or glued back together,44 or pinned to a board31 to prevent distortion during formalin fixation.9 10 17 Alternatively, a 14 gauge or purpose designed45 needle, skin punch biopsy needle or cork borer is used to obtain small pieces of fresh tissue. Resultant defects can be filled with ink to aid histological identification, or cork to prevent tissue distortion.9 10 17 32 Another technique is to incise a “peel” comprising surface prostatic tissue and underlying tumour, which is serially sectioned perpendicular to the inked surface and examined histologically to assess margin status and EPE. Underlying tumour is then removed and used for research.9 10 17 Epstein et al10 17 and Montironi et al9 describe frozen section examination of alternate posterior blocks from an area of tumour to assess margin status and EPE, with remaining tissue being used for research. Assessment of margin status in this context can be difficult as extraprostatic fibromuscular tissue is often flimsy and difficult to keep intact when sectioning. Once obtained, fresh tissue is generally snap-frozen (eg, in liquid nitrogen) stored at −70 to −80°C, and retained until the final pathology report has been issued in case it is needed for diagnostic purposes.

HANDLING PLA SPECIMENS

Pelvic lymph node (PLN) metastases confer a 50–70% risk of developing distant metastases within 5 years.29 Downward stage migration, owing to earlier detection and better patient selection for surgery, has produced an apparent decrease in the incidence of positive PLNs at the time of RP from 8.5% in 1988, to 1.2% in 1999.10 Consequently, and in view of the morbidity of the procedure, PLA is undertaken only in those patients at high preoperative risk of positivity, such as those with a serum PSA greater than 10, Gleason grade 4 in any biopsy, and/or more than three positive biopsies.46

There are currently no established guidelines for handling PLA specimens, although it is recognised that all grossly identifiable nodes should be submitted for histological examination.10 Studies suggest that accuracy of PLA staging is dependent on the number of lymph nodes examined. Barth et al observed the frequency of lymph node metastases in cases where 13 or more lymph nodes were examined at PLA to be double that found in cases where fewer nodes were examined.47 Weingärtner et al concluded that a PLA specimen should only be considered adequate and representative if approximately 20 nodes are recovered from the specimen.48 It has also been suggested that the prognosis of patients with lymph node metastasis is dependent on the total number of nodes involved by tumour.49 50 Pelvic nodes are often extensively replaced by adipose tissue, leaving only a small residual rim of lymphoid tissue, and estimation of node numbers in PLA specimens is often inaccurate.51 In our opinion, there is insufficient evidence in the literature to mandate a definitive assessment of node numbers, both total and positive, in PLA specimens. Since 6.5% of metastases are found in adipose tissue rather than readily identifiable nodes,53 some pathologists submit the entire PLA specimen, rather than harvesting individual nodes, for histological examination.10 In view of the very low frequency of nodal metastasis in contemporary literature, this comprehensive practice is probably unnecessary and might be better restricted to certain “high risk” patients.10 Local discretion should be exercised in this regard.

Examining each node through several levels and using immunohistochemistry for PSA or cytokeratins purportedly enhances detection of metastases in “low risk” patients from 5% to 11%;52 however, the clinical significance of nodal micrometastases identified by these techniques is uncertain and their routine use not recommended. Sentinel lymph node mapping in prostate cancer has shown potential but has a number of limitations and its utility remains uncertain.46

Frozen section analysis of PLA specimens

Intraoperative examination of PLNs is controversial. PLN metastasis is uncommon and, when present, is often small and easily missed by frozen section examination, with positive nodes only being subsequently identified in deeper paraffin sections of the cryostat block.54 Frozen section analysis is time consuming, expensive, has a low pick-up rate for PLN metastasis (only 2% in a recent study54), a high rate of false negativity (only two out of three positive nodes will be identified34), and should be only be undertaken if the result is likely to have immediate influence on the operative procedure. From a practical perspective, frozen section analysis of PLA specimens, often composed largely of adipose tissue, can be challenging and decent sections difficult to obtain. It is impractical to examine all tissue by frozen section and a selective approach is warranted (eg, freezing two or three small, firm nodes from each side, as advocated by Epstein et al,10 17 identifies more than 50% of small metastases). Harnden and Parkinson advocate sampling abnormal nodal areas together with a representative section from each node34 and examine between four and six sections from each block, taking further levels if original sections are suggestive but not diagnostic of carcinoma. Regardless of the method employed, ex cryostat and residual tissue should be fixed, processed and microscopically examined along with the main specimen.

Recommendations

• All grossly identifiable PLNs should be submitted for histological examination. Submission of residual tissue is left to local discretion.

• Routine examination of multiple levels and immunohistochemistry to detect micrometastases in PLNs is not recommended.

• Routine intraoperative frozen section examination of PLNs is not recommended.