SEARCH

SEARCH BY CITATION

Keywords:

  • prostate cancer;
  • staging;
  • lymph node metastases;
  • step section analysis;
  • immunohistochemistry

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Study Type – Therapy (case series)

Level of Evidence 4

What's known on the subject? and What does the study add?

Metastatic spread to the regional lymph nodes (LNs) is the single worst predictor of survival in prostate cancer. Knowledge of the LN status is crucial, as the treatment strategy is substantially altered in non-organ-confined disease. Current routine pathological protocols for evaluation of LN resection specimens do not demand a minimum of cross-sections per LN to be examined. Depending on the LN size, usually one to two sections are examined. This might lead to underestimation of the true metastatic burden.

The present study shows that additional examination of cross-sections in pelvic LNs and applying prostate cancer-specific cytokeratine immunostaining does not lead to significantly increased detection of prostate cancer metastases. However, the work-load and the expenses were significantly higher compared with routine evaluation.

OBJECTIVE

  • • 
    To evaluate the diagnostic gain in the detection of lymph node (LN) metastases of prostate cancer and the additional expenses of histological step-section analysis, including immunohistochemistry compared with routine histopathological evaluation.

PATIENTS AND METHODS

  • • 
    In a prospective study, 19 patients with prostate cancer at high risk of LN metastases (>cT2c and/or PSA level of >20 ng/mL and/or Gleason score >8) underwent sentinel-guided LN resection.
  • • 
    All palpable LNs were submitted to step-section analysis in 200-µm sections and concomitant immunohistochemical staining for cytokeratine (AE1/AE3), in addition to routine histopathology of one or two haematoxylin and eosin-stained sections per LN.
  • • 
    The number of positive LNs and LN-positive patients for each method was compared; additional expenses in labour time and material for the extended evaluation were estimated.

RESULTS

  • • 
    In all, 413 LNs were resected; 220 LNs were palpable and were included in the study.
  • • 
    In seven of the 19 patients routine histopathological evaluation revealed LN metastases in 24 of 220 LNs (10.9%).
  • • 
    Extended LN evaluation with step sectioning and cytokeratin immunohistochemistry did not reveal any additional patients with LN metastases.
  • • 
    In one patient already diagnosed with LN metastases on routine histology, four additional LN metastases were detected upon extended LN evaluation. Three LNs of two patients, one of them pN0, contained disseminated tumour cells.
  • • 
    Compared with conventional histological evaluation, serial-section analysis and immunohistochemistry increased expenses in materials and labour time 18.7-fold.

CONCLUSIONS

  • • 
    Serial-section analysis seems to have only a minimal diagnostic gain; however, valid conclusions cannot be drawn, as not all LNs were submitted to extended evaluation.
  • • 
    Considering the additional expenses, extensive LN evaluation in prostate cancer cannot generally be recommended.

Abbreviations
LN

lymph node

PLND

pelvic LN dissection

H&E

haematoxylin and eosin

RRP

retropubic radical prostatectomy

ADASP

Association of Directors of Anatomic and Surgical Pathology

NHT

neoadjuvant hormone deprivation therapy

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Metastatic spread to regional lymph nodes (LNs) in prostate cancer is significantly associated with disease progression and death from prostate cancer [1–3]. To date, pelvic LN dissection (PLND) is considered the ‘gold standard’ for staging, as reliable imaging methods do not exist [4]. Several methods have been discussed to improve the accuracy of LN sampling during the procedure, including extended field PLND [4,5] or sentinel LN-guided PLND [6,7]. Generally, staging accuracy is not only determined by LN sampling but also pathological LN evaluation significantly influences the detection rate of LN metastases [8]. Thorough palpation and meticulous dissection of the resection specimen to isolate single LNs or defatting of the connective tissue to identify the LN tissue contributes to an effective LN sampling [9]. In addition, the detection rate of metastases may depend on the number of sections examined per LN [10]. Conventional histopathological evaluation is usually limited to few cross-sections per LN and therefore might not detect LN metastases with a smaller diameter and may consequently contribute to an understaging in a substantial number of patients. Routinely, haematoxylin and eosin (H&E) staining is applied for the investigation of LNs [11]. Immunohistochemical staining for cytokeratine is more sensitive and may facilitate the detection of low-volume metastases and thus, further increase the sensitivity of the staging procedure. The putative diagnostic gain of an extended evaluation of the LN resection specimens by increasing the number of cross-sections and applying immunohistochemistry has to be weighed against the consequent increase of expenses in material and labour time. Moreover, the prognostic importance of micrometastases or even disseminated tumour cells detected by more sensitive diagnostic methods still remains a matter of debate [7].

The goal of the present study was to prospectively investigate the sensitivity and clinical feasibility of serial-section analysis and immunohistochemistry of the complete LN in resection specimens of patients at increased risk of advanced prostate cancer compared with routine histopathological examination. An additional goal was to analyse the incremental costs of the extended evaluation.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The LNs of 19 patients with histologically confirmed prostate cancer and either suspicious LNs on CT or MRI or high-risk prostate cancer according to the D'Amico classification (clinical stage >cT2 and/or PSA level of >20 ng/mL and/or biopsy Gleason score >8) [12] were included in the study. Patients underwent either laparoscopic sentinel LN-guided staging LN resection or retropubic radical prostatectomy (RRP) with sentinel guided PLND. In all, 413 LNs were resected; 220 of these could be identified by palpation and were included in the study.

The study was authorised by the Ethics Committee of the University of Tuebingen (No. 151/2006V); all patients provided written informed consent.

Sentinel LN-guided laparoscopic staging lymphadenectomy was carried out as described before [6,7]. All dissected lymphadenectomy specimens were sampled separately considering the anatomical locations of sentinel LN areas of the prostate.

For processing of the lymphadenectomy specimens and histological examination, the resection specimen was thoroughly examined and all palpable LNs were separated. A representative fraction of each LN was separately snap-frozen in liquid nitrogen and stored at −80 °C for further analysis. Each palpable LN was then inserted in an individual cassette.

LNs with a diameter of >7 mm were bisected longitudinally along the greatest dimension. Both halves were inserted side-by-side into the cassettes. In cases where both halves exceeded a depth of 7 mm or the longest dimension did not fit into a routine cassette, LNs were distributed into two cassettes by further bisecting the two halves. The remaining tissue was completely embedded and investigated routinely by H&E staining. Specimens were fixed in 4.5% formaldehyde overnight and embedded in paraffin. All palpable LN specimens were completely serially sectioned at 200-µm intervals. The section with the greatest LN diameter was processed according to a routine protocol with H&E staining. If the thickness of the embedded LN tissue was >5 mm, one additional section was processed according to the routine H&E staining protocol. Consequently, at least two cross-sections were processed by H&E staining in LNs >7 mm and there was at least one cross-section in LNs with a smaller diameter. All other cross-sections were immunohistochemically stained. For immunostaining, antibodies against cytokeratin AE1/AE3 (Dako, dilution 1:50) were used and detected by a secondary antibody applying the iVIEW 3,3′-diaminobenzidine (DAB) detection kit (Ventana Medical Systems, Tucson, USA). Routine pathology and step-section analysis were carried out by a single senior pathologist (K.S.).

For the definite pathology report, all LN sections were investigated for presence of metastases, micrometastases (greatest dimension <0.2 mm) or disseminated tumour cells according to the classification proposed earlier by the International Union Against Cancer [13]. Step-section analysis with immunohistochemistry was considered the ‘gold standard’ for the definition of the LN status (pN0, pN1). LNs with no evidence of metastases in routine pathology but positive findings in step-section analysis and immunohistochemistry were considered LN positive (pN1). The current seventh edition of the Union Internationale Contre le Cancer TNM classification of malignant tumours does not recognise the designations ‘micrometastasis’ or ‘disseminated tumour cell’[14], therefore LNs with micrometastases or disseminated tumour cells only, were considered as pN0, also in accordance with the current recommendations of the Association of Directors of Anatomic and Surgical Pathology (ADASP) and the Royal College of Pathology (RCPath) [9,11,15].

For the cost analysis, calculations were based on the mean number of histological sections per LN and mean number of LNs per patient. The costs for materials, e.g. microscopic slides and staining antibodies, were based on the pricing information of the manufacturers, expenses for labour time were calculated based on the listings of the Collective wage agreement for German state employees (BAT, see http://www.medizin.uni-tuebingen.de/uktmedia/Forschung/PDF_Archiv/Kostenstze_IZKF_2009_2010_Jun10-p-14098.doc).

The data were analysed using the statistical software package JMP® (SAS Institute, Cary, NC, USA). Individual percentages of positive LNs were compared by anova. The cost analysis of LN evaluation was compared by Wilcoxon-Kruskal-Wallis tests. A significance level of P < 0.05 was considered statistically significant.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

In all, 15 patients underwent sentinel LN-guided laparoscopic staging LN resection, and four primary RRP. The mean (median, range) age was 65.7 (67, 51–77) years, the median (range) Gleason sum was 7 (5–9), the mean (median, range) pre-treatment serum PSA level was 42.2 (23.0, 3.2–154.0) ng/mL. The mean (median, range) number of resected LNs was 21.7 (20, 6–44) per patient.

EXTENDED LN EVALUATION

In all, 413 LNs were resected and 220 of these LNs could be identified by palpation and were submitted for serial-section analysis in 200-µm sections (mean 12.0, median 12, range 4–23). The mean (median, range) greatest dimension of the palpable LNs was 16.4 (15, 6–55) mm, with the largest LN measuring 55 × 10 × 5 mm. A mean (median, range) of 1.3 (1, 1–2) routine sections per LN were analysed according to the routine H&E protocol. The mean (median, range) number of sections from serial-section analysis in addition to the slides stained according to the routine H&E protocol was 22.9 (22, 12–41) per LN, resulting in a mean of 274.8 additionally performed sections per patient.

HISTOPATHOLOGICAL RESULTS

Routine histopathological analysis revealed metastatic LN involvement in seven of the 19 patients (pN1, 36.8%). Serial-section analysis of the H&E-stained slides did not detect additional patients with metastatic disease. The median (range) number of positive LNs in pN1 patients was 2 (1–8) for routine histology and 4 (1–8) for the extended evaluation (P= 0.74). On a patient-by-patient basis, the mean number of positive LNs did not differ significantly between conventional evaluation and serial-section analysis (individual mean: conventional 11.6% [min-max 0.8–22.5] vs serial sectioning 13.1% [min-max 2.3–24.0], P= 0.84).

On a LN-by-LN basis, routine histopathological evaluation with H&E staining identified metastases in 24/220 LNs (10.9%) and step-section analysis with anti-cytokeratine immunohistochemistry revealed metastatic disease in 28/220 LNs (12.7%, not statistically different, P= 0.84). All LNs that were additionally detected in serial sectioning and immunohistochemistry were found in one patient. However, in this patient the LN status did not change, as metastases had already been detected in two LNs by routine histopathology. The mean (median, range) extension of the metastases was 17.5 (16.5, 2–35) mm. Single disseminated tumour cells were identified by step sectioning in one additional LN of an LN-positive patient and in two LNs of another patient with no evidence of LN metastases (pN0, Fig. 1).

image

Figure 1. Overview of the histological results for a single LN. DTC, disseminated tumour cells.

Download figure to PowerPoint

EXPENSE OF EXTENDED LN EVALUATION

To evaluate the feasibility of the extended LN evaluation, the additional expenses in material and labour time were analysed. The following calculations are based on the mean number of sections per LN (1.3 sections/LN for routine analysis vs 24.3 sections/LN for step-sectioning analysis; 24.3/1.3 = 18.7-fold). Therefore, step sectioning of the paraffin-embedded specimens carried out manually on a rotating microtome by a technician expanded the work-load 18.7-fold. Similarly, expenses for microscopic slides and cover slides were 18.7-fold. The additional immunohistochemical staining procedures were carried out in an automated staining machine with negligible additional effort. However, expenses for primary and secondary antibodies have to be taken into account. Based on the mean of 12.0 palpable LNs per patient, the mean number of sections processed and analysed by step sectioning and immunohistochemistry was significantly higher compared with conventional pathological evaluation (274.8 [min-max 225.2–324.4] vs 15.6 [min-max 12.4–18.8], P < 0.001). Consequently, expenses for the extended evaluation were significantly higher (1321.97 €/patient [min-max 1181.54–1462.40] vs 60.46 €/patient [min-max 51.55–69.37], P < 0.001). An overview of the estimated costs for conventional and extended LN evaluation is given in Table 1.

Table 1.  Estimated expenses for LN evaluation
VariableUnit priceEstimated costs per patient P
ConventionalExtended
Sections/patient    
 N (min-max) 15.6 (12.4–18.8)274.8 (225.2–324.4)<0.001*
Material    
 Slides + cover slides0.10 €/section15.6 × 0.10 = 1.56 €274.8 × 0.10 = 27.48 € 
 Immunohistochemistry2.00 €/section274.8 × 2.00 = 549.60 € 
Labour time    
 Technician    
  Labour time0.08 h/section15.6 × 0.08 = 1.25 h274.8 × 0.08 = 22.90 h 
  Costs23.60 €/h†≥29.45 €≥540.44 € 
 Pathologist    
  Labour time0.02 h/section15.6 × 0.02 = 0.31 h274.8 × 0.02 = 5.50 h 
  Costs37.20 €/h†≥11.61 €≥204.45 € 
Total    
 € per patient (min-max) 60.46 (51.55–69.37)1321.97 (1181.54–1462.40)<0.001*

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Pagliarulo et al. [16] showed that patients with clinically occult metastases missed at the time of PLND have the same risk of biochemical recurrence and death from prostate cancer as patients with metastases detected at the time of surgery. Still, the prognosis of patients with LN involvement is not invariably poor. Messing et al. [17] showed that the immediate onset of androgen deprivation prolongs the interval to biochemical progression and provides survival advantages in patients diagnosed with LN metastases at the time of surgery compared with a deferred onset of therapy. For this reason, exact LN staging is of utmost importance to the patient as the knowledge of the LN status significantly alters the therapeutic strategy.

Various investigations have shown that staging accuracy widely depends on the extent of PLND [18–20]. However, not only LN sampling but also pathological processing, particularly the number of sections examined per LN, may affect the detection rate of metastases. In addition, immunohistochemical staining of the sections may facilitate the detection of low-volume metastatic disease and further improve staging accuracy. However, to date there has been no consensus about the number of sections per LN to be examined; the ADASP does not state a minimum of sections and routine application of immunohistochemistry is not advised [15]. In the present prospective investigation, we aimed to determine the difference in detection rate of routine histopathology with H&E staining and serial-section analysis with immunohistochemical staining for cytokeratine.

The patients in the present study population were clinically at high risk of LN metastases; this fact might explain the high rate of patients with metastatic disease (seven of 19). Interestingly, serial section analysis in 200-µm slices and immunostaining did not change the LN status in any of these patients. Only four additional LNs in one patient revealed metastatic disease after extensive processing and meticulous analysis. Similar observations have been made by other groups. In a study of 32 patients with organ-confined prostate cancer and no evidence of LN involvement in routine H&E histopathology, immunocytochemical staining for a pan-cytokeratin monoclonal antibody detected only one additional patient with micrometastatic disease in one LN [21]. In another study in 41 patients with different stages of disease staining for PSA-antigen, prostate-specific acidic phosphatase and cytokeratine detected only two additional patients with occult LN metastases [22]. In a more recent study, Martinéz-Piñero et al. [23] investigated 154 patients undergoing PLND for prostate cancer. Nine of the patients (5.8%) had evidence of LN metastases in routine H&E staining. Immunohistochemistry for PSA and pan-cytokeratine (AE1/AE3) did not detect any additional LNs that were not identified by routine histopathology. Arguably, the detection rate of the aforementioned studies might be low because only a limited number of sections of each LN were investigated. In an extensive study in 194 patients undergoing RP and sentinel LN-guided PLND Wawroschek et al. [24] submitted all sentinel LNs to serial-section analysis with sections at 400-µm intervals and immunostaining with a pan-cytokeratine antibody. Even the combination of serial sections and immunostaining improved the detection rate of metastases only marginally from 22.2% in conventional histopathology with H&E and step-section analysis in 2-mm sections to 26.8% with immunohistochemistry and serial sections. In contrast, in a study of 42 patients Fukada et al. [25] showed an increase in the detection rate from 19.0% to 31.0%, if LNs had been step sectioned in 250-µm intervals and stained for anti-cytokeratine (AE1/AE3). However, when these patients were differentiated into groups with and without neoadjuvant hormone deprivation therapy (NHT), it became evident that all additionally detected patients belonged to the group under NHT and none of the hormone-naive patients harboured occult LN metastases. The authors concluded that hormone deprivation might obscure the common histopathological findings of metastatic disease in prostate cancer and immunohistochemistry might facilitate the diagnosis in this particular group of patients. Moreover, the patients with NHT were probably at higher risk of LN metastases, increasing the likelihood of detecting additional metastases in a more extensive evaluation.

The additional preparation of the LN in the present study extended the pathological work flow significantly. The estimated labour time was 18-fold compared with routine histopathological reporting. Additional costs for labour time and material, e.g. antibodies and microscopic slides, summed up to an average of €1321.97 per patient, 23-fold that of conventional evaluation (P < 0.001, Table 1).

Considering the significant increase in expenses and the marginal diagnostic gain of step-section analysis, less complex methodological steps to increase the diagnostic accuracy should be considered. Several steps have been discussed by the International Society of Urological Pathology (ISUP) [26]: embedding not only of palpable LN, but also of the fatty tissue may increase the LN yield. This was confirmed in the present study: a little over half of the LNs (220/413) could be identified by palpation, the remainder of the LNs were detected after embedding the fatty tissue surrounding the LNs. Fat cleavage by xylene after dehydration of the specimen has been shown to aid in identifying single LNs and therefore improving LN yield in colonic cancer [27]. Certainly, diligent processing of the LN samples is a key step in accurate LN staging; however, there is no debate that all elaborate technical adjustments to the histopathological evaluation of the resection specimens will not replace meticulous, extended PLND.

One limitation of the present study is that only palpable LNs were submitted to serial-section analysis. Consequently, the total number of LN included in the study and therefore the number of potentially positive LNs was decreased compared to the whole resection specimen. However, the goal of the present study, was to examine the detection rate as a function of pathological evaluation on a LN-by-LN basis. Therefore, the extent of the LN resection should not have influenced the results of the analysis (except for the total number of LNs and potential number of positive LNs). Another limitation of the present study is the low number of patients enrolled. However, based on the number of LNs examined we think that valid conclusions can be drawn.

Taking the present investigation and the aforementioned studies into account, it seems that conventional histopathology with one to two sections per LN, depending on the LN diameter, and staining with H&E indeed diagnoses most metastases. The additional evaluation with step-section analysis and immunohistochemistry does not significantly improve detection of tumour spread into regional LNs. The detection of micrometastases and disseminated tumour cells may necessitate the processing of LN specimens with a resolution beyond 200 µm. Recently, we proposed a novel method to monitor even single disseminated tumour cells in a LN [28]. By calculating the tumour cell density, we proposed a novel parameter to measure the true tumour cell burden. Applying this method prospectively might contribute to a better understanding of ‘sub-microscopic’ LN tumour dissemination in prostate cancer and may help to define more precisely whether the current routine evaluation is sufficient to evaluate the LN status of patients with prostate cancer. Still, the question whether a quantitative analysis of the true LN tumour burden is of real clinical and prognostic significance requires prospective validation within larger patient cohorts and longer follow-up.

In conclusion, in the present patients with a high risk of non-organ-confined prostate cancer undergoing modified sentinel guided LN resection, we detected LN metastases in about a third of the patients. Serial step-section analysis in 200-µm intervals with concomitant immunohistochemical staining for pan-cytokeratine, improved the detection rate only marginally from 10.9% to 12.7% of the palpable LN in the study. Because only palpable LNs rather than the whole template were submitted to serial sectioning and immunohistochemistry, valid conclusions on the diagnostic value cannot be drawn. However, the extended evaluation resulted in a considerable increase in labour time and expenses. Based on the findings of the present study, extensive LN evaluation in prostate cancer cannot be recommended.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The authors thank Miriam Germann for critically reading the manuscript.

CONFLICT OF INTEREST

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

None declared. Source of funding: This work was generously supported by the Program for Applied Clinical Research of the Eberhard-Karls-University Tuebingen (AKF-Grant #187-0-0).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES