Sentinel lymph node biopsy in rectal cancer—Not yet ready for routine clinical use

Original communications Sentinel lymph node biopsy in rectal cancer—Not yet ready for routine clinical use A. Bembenek, MD, B. Rau, MD, T. Moesta, MD, J. Markwardt, MD, C. Ulmer, MD, S. Gretschel, MD, U. Schneider, MD, W. Slisow, MD, and P. M. Schlag, MD, Berlin, Germany

Background. The value of sentinel node biopsy in visceral cancers is uncertain. We evaluated the feasibility and utility of radiocolloid lymphatic mapping and selective lymph node sampling in patients with rectal cancer. Methods. Forty-eight patients with rectal cancer were investigated. Thirty-seven patients had already undergone preoperative radiochemotherapy for locally advanced tumors. Eleven patients underwent primary surgery. An endoscopic injection of 1 mL technetium 99m–sulfur-colloid into the peritumoral submucosa was performed 15 to 17 hours before surgery. Ex vivo identification of the nuclide-enriched ‘‘sentinel lymph nodes’’ (SLNs) was performed using a hand-held c-probe. The selected SLNs were then carefully and systematically examined using serial sections and immunohistochemistry. Results. One or more SLNs were found in 46 of the 48 patients. The SLN detection rate was 96%. Sixteen of the 48 patients had lymph node metastases (35%). In 7 of the 16 patients, the SLNs correctly represented the nodal status. In 9 of the 16 patients, the SLN was tumor-free whereas non-SLN harbored metastases. This result represents a sensitivity of only 44%, and a false-negative rate of 56%. Further analysis showed that the method correctly predicted the nodal status only in the small subgroup of 5 patients with early cancer without preoperative radiation. In 4 patients, juxtaregional lymph nodes were excised on the basis of intraoperative radiocolloid detection, leading to upward staging in 1 patient. Conclusions. Sentinel lymph node biopsy using the radiocolloid technique with ex vivo lymph node identification shows a relatively high detection rate; however, the sensitivity in patients with locally advanced/irradiated rectal cancer is low. Nevertheless, the detection of juxtaregional metastases can improve staging in some patients. Further studies should focus on patients with early rectal cancers where the data were more promising. (Surgery 2004;135:498-505.) From the Departments of Surgery and Surgical Oncology, Pathology, and Nuclear Medicine, Charite´ Campus Buch, Humboldt-University and Robert-Ro¨ssle-Klinik in ‘‘Helios Klinikum Berlin,’’ Berlin, Germany

LYMPHATIC MAPPING AND SENTINEL LYMPH NODE BIOPSY (SLNB) are of value in the evaluation of the nodal status in patients with breast cancer and melanoma. The method predicts those patients with the highest probability of metastatic disease.1-3 However, the value of SLNB in patients with visceral cancers is uncertain. Preliminary studies suggest that lymphatic mapping is feasible in patients with gastroAccepted for publication October 10, 2003. Reprint requests: Peter M. Schlag, MD, Department of Surgery and Surgical Oncology, Charite´, Campus Berlin-Buch, RobertRoessle-Clinic in ‘‘Helios-Klinikum Berlin’’, Lindenbergerweg 80, 10437 Berlin. 0039-6060/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.surg.2003.10.004

498 SURGERY

intestinal cancers.4-12 Data on SLNB in colorectal cancer are limited and contradictory.7-12 Furthermore, few patients with rectal cancer were included in these studies. We conducted a prospective feasibility study to assess the predictive value of SLNB in patients with rectal cancer. We reasoned that nodal staging would improve selecting patients for adjuvant therapies and would facilitate the detection of lateral pelvic lymphatic drainage. The latter may serve to identify patients who could benefit from extended lymph node dissection. Because of the close vicinity of the pelvic region and the narrow space around the mesorectal tissue, a subserosal injection of dye followed by intraoperative tracing appeared impracticable. Instead, we used a radiocolloid method providing

Surgery Volume 135, Number 5

Fig 1. Preoperative lymphoscintigraphy. Thin arrows indicate SLNs along mesenteric vessels; thick arrow indicates tumor site.

a relatively long period of tracer deposition in the lymph nodes. This approach enabled us to scan for residual activity in vivo using a hand-held c-probe, as well as to identify SLNs from the specimen ex vivo. PATIENTS AND METHODS Patients. Forty-eight patients (14 women, 34 men) with histologically proven and locally confined rectal cancer were enrolled. The pretherapeutic staging was based on endoscopic ultrasonography and was completed by computed tomography (CT) or magnetic resonance imaging (MRI) of the pelvic region, abdominal ultrasonography, and chest roentgenography. Forty-three patients were thereby staged as uT3/4, and 5 as uT1/2. Thirty-seven patients underwent neoadjuvant radiochemotherapy for locally advanced cancer (endosonographically uT3/4), whereas 11 patients underwent primary surgical therapy for cancer with infiltration limited to the muscle layer (uT2, n = 5) or for locally advanced cases with contraindications for neoadjuvant radiation (n = 6) (Table I). Surgical procedures were 32 anterior resections, 5 anterior resections with low

Bembenek et al 499

colo-anal anastomosis preserving the anal sphincter, and 11 abdomino-perineal extirpations. Methods. Our institutional review board approved the study. After written informed consent was obtained, 1 mL technetium 99m–sulfur-colloid (Nanocis, particle size: 100 lm; Schering, Berlin, Germany) was injected endoscopically into the submucosal space immediately adjacent to the tumor. In tumors enabling the passage of the flexible endoscope, the total dose was administered in 4 portions proximally, distally and to both sides laterally from the tumor. In obstructing tumors, the oral injection site was not attainable, and the dose was thus administered in only 3 portions. The injection was done 16 to 18 hours before surgery. One hour before operation a radiocolloid-scintigraphy was recorded with the c-camera to provide a general impression about the radionuclide enrichment in the lymph nodes (Fig 1). Immediately after resection, the residual in vivo activity at the remaining lymph node regions in the operative field was examined with the c-probe. When a residual radionuclide-enriched lymph node was encountered, the node was excised and submitted for separate histopathologic examination. The sentinel lymph nodes (SLNs) were detected by examining the specimen ex vivo with a handheld c-probe. Detection with the c-probe was guided by specimen scintigraphy and a coordinate system underneath the specimen that corresponded to a coordinate system projected on the scintigraphic picture. This technique enabled an exact correlation between the ‘‘hot spot’’ on the scintigraphic picture and the corresponding lymph node in the specimen identified by the hand-held c-probe (Fig 2). The first 15 patients enrolled had their specimens prepared without mesorectal separation by the pathologist. Thereafter, we took a precaution to avoid interference between the radiocolloid injection site around the tumor and the closest pararectal lymph nodes. The pathologist separated the mesorectal tissues, including the lymph nodes and the inferior mesenteric vessels, from the bowel wall before performing the specimen scintigraphy and the preparation of the lymph nodes (Fig 3). We termed the evaluation of the first 15 specimens without mesorectal separation ‘‘phase 1.’’ The 33 specimens evaluated subsequently were termed ‘‘phase 2.’’ The positions of the radionuclideenriched lymph nodes were registered on the coordinate system. Histopathology. The routine assessment included several steps. Lymph nodes up to 10 mm in diameter were bivalved. The larger lymph nodes

500 Bembenek et al

Surgery May 2004

Fig 2. Specimen scintigraphy (A) and resected specimen (B) including tumor and mesorectum on coordinate system. Thin arrows indicate SLNs along mesenteric vessels; thick arrow indicates tumor site.

Table I. Results according to postoperative tumor stage No. patients

SLN detection rate

Nodal positive patients

SLN/Patient (Median)

Sensitivity

Negative predictive value

ypT0 ypT1 pT2 ypT2 pT3 ypT3

3 1 5 17 7 15

3/3 1/1 4/5 (80%) 17/17 (100%) 6/7 (86%) 15/15 (100%)

0 0 2 3/17 (18%) 3/7 (43%) 8/15 (53%)

7 7 4 2.5 4 4

— — 2/2 0/3 0/3 4/7

— — — — — —

Overall

48

46/48 (95%)

16/46 (35%)

7/16 (44%)

0.77 (77%)

4

ypT, Tumor stage after preoperative radiochemotherapy.

were cut in 2- to 3-mm sections and processed to paraffin blocks. After hematoxylin and eosin (H&E) staining, the diagnosis was determined. If the processed SLN(s) did not reveal metastasis, the SLN underwent stepwise sections in levels of 250 lm until the SLN was sampled completely. At each level, serial sections were cut at 5 lm thickness

for at least 1 slide undergoing H&E and 1 slide undergoing immunohistochemical staining (MNF 116, DAKO, Hamburg, Germany). With this approach, metastases with a maximum size of 0.25 mm were identified with 100% probability. With a maximum size of 1 mm this value was reduced to a 50% probability.13

Surgery Volume 135, Number 5

Bembenek et al 501

Fig 3. A and B, Mesorectum on coordinate system after separation from tumor. Thin arrows indicate SLNs along mesenteric vessels; thick arrow indicates tumor site. (Figure continued)

The number of resected lymph nodes and the number of resected SLNs before and after separation of the mesorectal tissue were compared by the Mann-Whitney test for independent variables. P < .05 was accepted as significant. RESULTS SLNs were identified in 46 of the 48 patients. The SLN detection rate was 96%; 16 of the 46 patients had histologically positive lymph nodes (35%). In 7 of these 16 patients, one or more of the SLNs contained metastatic disease correctly indicating the positive nodal status. Nine patients had lymph node metastases in non-SLNs in spite of the fact that the respective SLNs were tumor-free, representing a false-negative result. The resulting overall sensitivity for the detection of a positive nodal status was 7/16 (44%) (Table II). This result corresponded to a negative predictive value of 77%. The overall correlation between the SLN status with the definitive nodal status was 37/46 (80%).

The median number of lymph nodes examined per patient was 18 (range, 9 to 69). The ex vivo separation of the mesorectal lymphatic tissue from the bowel wall before the specimen scintigraphy, the procedure performed after patient 15 (phase 1), increased the median number of examined lymph nodes from 14 to 19 (P < .01). The median number of SLNs examined increased numerically, but not significantly, from 3 to 4 (P = .20). However, the sensitivity was not increased. This value was 40% in the phase 1 patients and 36% in the phase 2 patients (ex vivo separation of the mesorectal lymphatic tissue from the bowel wall) (Table III). To evaluate the influence of the histopathologic tumor stage, we compared SLNB sensitivity in patients with postoperative tumor stage (y)pT0-2 to patients with stages (y)pT3-4. The sensitivity was similar in both groups (40%). Similarly, preoperative neoadjuvant radiochemotherapy was not correlated with a difference in SLNB sensitivity. The sensitivity was 50% with preoperative

502 Bembenek et al

Surgery May 2004

Fig 3. (continued) C, Ex vivo SLN identification at specimen with c-probe.

Table II. Results before (phase 1) and after (phase 2) ex vivo separation of mesorectal tissue from the rectal wall to avoid overlapping radioactivity Phase 1 Phase 2

No. patients

Detection rate

Nodal positive patients

SLN/Patient (Median)

Sensitivity

13 35

100% 13/13 94% 33/35

4 12

3 4

(50%) 2/4 (42%) 5/12

radiochemotherapy versus 42% without this treatment. Only the small group of patients (n = 5) with early tumor stages (uT2) who did not undergo preoperative chemoradiation revealed a correct identification of the nodal status by SLNB in all patients (Table II). This small group included 2 node-positive patients. Immunohistochemistry did not reveal any additional metastases in the SLNs examined. In 4 patients, a juxtaregional SLN underwent biopsy during the operative procedure. Three patients had an SLN in the interaortocaval space and 1 in the left parailiac region. In 1 of these patients who had a local staging of ypT0N0M0, the excised lymph node contained a site of metastasis.

This finding caused this patient to be upstaged to pM1 status (Fig 4). The patient later had development of pulmonary metastases. The remaining 3 juxtaregional lymph node excisions were done in 1 patient staged as ypT2N0G2M0 and 2 patients staged each as ypT2N1G2-(23) tumors. DISCUSSION Our results demonstrate that SLNB with the radiocolloid technique is feasible in patients with rectal cancer. However, several problems occurred with this disease that may make the condition unique in terms of SLNB interpretation. In contrast to melanoma or breast cancer, the primarily draining lymph nodes are in close proximity to the

Bembenek et al 503

Surgery Volume 135, Number 5

Fig 4. Radionuclide enrichment in interaorto-caval space and corresponding lymph node in CT scan.

Table III. Overview of lymphatic mapping and selective lymph node biopsy in colon cancer

Joosten ’998 Merrie ’0121 Wong ’0120 Saha ’0110 Bilchik ’0118 Paramo ’0222 Feig ’0216 *Kitagawa ’0212 Trocha et al ’0323

No. patients

Mapping Technique (time point of injection)

Time point of identification

50 26 26 203 126 55 48 56

BD (intraoperative) BD + RT (intraoperative) BD (postoperative) DT DT BD (intraoperative) BD (intraoperative) RT (preoperative)

48

FT + RT (intraoperative)

Ex vivo Ex vivo Ex vivo Intraoperative Intraoperative Intraoperative In vivo Intraoperative (c-probe) In vivo

Nodal positive patients

Detection rate

Sensitivity

Upstaging by IHC

57% 30% 46% 37% 35% 33% 33% 39%

70% 88% 92% 98% 97% 82% 98% 91%

40% 55% 92% 90% 92% 91% 62% 81%

13% 25% 20% 11% 8% 17% 13% n. s.

38%

98%

84%

21%

RT, Radiocolloid technique; DT, blue dye technique; n. s., not stated.

*Including 60% of patients with rectal cancer.

tumor and the peritumoral radiocolloid injection site causes overlapping radioactivity. This anatomic situation renders identification of radionuclideenriched pararectal lymph nodes difficult. To solve the problem, we separated the mesorectal tissue ex vivo from the bowel wall before starting the search for SLNs after the fifteenth patient. This modification made the identification of selective pararectal SLNs technically easier. Furthermore, the number of lymph nodes identified was increased. However, in spite of the increased number of lymph nodes examined and the high detection rates, the

sensitivity of SLNB in the detection of metastatic lymph nodes did not increase further. As assumed for locally advanced breast tumors,1,14 our results indicate that an advanced tumor stage hinders successful SLNB in rectal cancer. The results also indicate that downstaging by preoperative irradiation and chemotherapy to a ypT2-tumor stage does not improve the sensitivity of SLNB in these patients. Whether or not the irradiation itself alters the functional lymphatic capacity remains speculative. Only in the small group of patients that presented initially with an

504 Bembenek et al

early tumor stage (uT2) and who did not undergo irradiation and chemotherapy, the SLNs correlated well with the final nodal status of the patients. In these patients, the only metastatic lymph node in both of the nodal positive cases was the SLN. However, this is a very small subgroup (n = 5), and any conclusions drawn must be preliminary. The currently available experience is contradictory and is based on studies of patients with ‘‘colorectal cancer.’’ Most of these tumors were actually colon cancer, and no or only a few patients with rectal cancer were included. Moreover, the methods applied varied markedly. For instance, Joosten et al8 performed intraoperative blue dye mapping. However, the postoperative SLN identification failed to reveal a satisfactory sensitivity for the prediction of the nodal status. In contrast, Saha et al10 and Bilchik et al15 used intraoperative blue dye mapping with subsequent intraoperative identification of the SLN. They described high detection rates, high sensitivities, and a significant percentage of micrometastasis in H&E-negative patients by serial sections and immunohistochemistry. The investigators also found aberrant lymphatic drainage in several patients leading to upstaging and an alteration of the extent of surgery.7 Other studies, however, could not confirm these results.11,16 Analyzing the available data for factors influencing the value of the SLNB, most studies describe worse results when patients with a high percentage of advanced tumors (T3/T4) were included.8,12,16-18 This was the case even with ex vivo identification of the SLNs after in vivo labeling.8,19 In agreement with this result, Kitagawa et al12 recently described a series of patients that included 60% of patients with rectal cancer. They observed false-negative results in T3-tumors, whereas the sensitivity for pT1/2-tumors was 100%.12 An alternative ex vivo detection technique performing the complete procedure on the resected specimen was described by Wong et al.20 They used postoperative submucosal injection of Lymphazurine 1% around the tumor after antimesenteric, longitudinal incision of the specimen in 26 patients with ‘‘colorectal’’ cancer. Their series included 24 patients with colon cancer, but only 2 patients with rectal cancer. They reported a detection rate of 92% and a sensitivity of 94%. The much higher sensitivity to detect metastases compared with in this study is in agreement with other studies of patients with colon cancer and may indicate that the lymph drainage from colon tumors is less complicated than from rectal cancers. From a technical point of view, the tracing and preparation of a blue channel after a manual massage in the

Surgery May 2004

specimen may enable the detection of lymph nodes that have already lost their function. We suggest that future studies should focus on patients with rectal cancer with locally limited disease (T1- and T2-tumors), as opposed to those who have more advanced disease, because SLNB appears more promising in patients with limited disease. Whether SLNB frequently fails in patients with rectal cancer is related to functional or anatomic reasons or is the result of advanced disease are important questions to be answered. REFERENCES 1. Borgstein PJ, Pijpers R, Comans EF, van Diest PJ, Boom RP, Meijer S. Sentinel lymph node biopsy in breast cancer: guidelines and pitfalls of lymphscintigraphy and gamma probe detection. J Am Coll Surg 1998;186:275-83. 2. Chu KU, Turner RR, Hansen NM, Brennan MB, Bilchik A, Giuliano AE. Do all patients with sentinel node metastasis from breast carcinoma need complete axillary node dissection? Ann Surg 1999;229:536-41. 3. Chu KU, Turner RR, Hansen NM, Brennan MB, Giuliano AE. Sentinel node metastasis in patients with breast carcinoma accurately predicts immunohistochemical detectable nonsentinel node metastasis. Ann Surg Oncol 1999;6:756-61. 4. Saha S, Wiese D, Badin J, Beutler T, Nora D, Ganatra BK, et al. Technical details of sentinel lymph node mapping in colorectal cancer and its impact on staging. Ann Surg Oncol 2000;7:120-4. 5. Bembenek A, Bayraktar S, Gretschel S, Ulmer C, Schulze T, Markwardt J, et al. Sentinel lymphonodectomy in gastrointestinal cancer-where are we now? Onkologie 2002;24:334-40. 6. Schlag PM. Sentinel Lymphknoten Biopsie. Landsberg: Ecomed Verlagsgesellschaft; 2001. 7. Tsioulias GJ, Wood TF, Morton DL, Bilchik AJ. Lymphatic mapping and focused analysis of sentinel lymph nodes upstage gastrointestinal neoplasms. Arch Surg 2000;135: 926-32. 8. Joosten JJA, Strobbe LJA, Wauters CAP, Pruszczynski M, Wobbes T, Ruers TJM. Intraoperative lymphatic mapping and the sentinel node concept in colorectal carcinoma. Br J Surg 1999;86:482-6. 9. Cserni G, Vajda K, Tarjan M, Bori R, Svebis M, Baltas B. Nodal staging of colorectal carcinoma from quantitative and qualitative aspects. Can lymphatic mapping help staging? Pathol Oncol Res 1999;5:291-6. 10. Saha S, Bilchik A, Wiese D, Espinosa M, Badin J, Ganatra BK, et al. Ultrastaging of colorectal cancer by sentinel lymph node mapping technique—a multicenter trial. Ann Surg Oncol 2001;8:94-8. 11. Broderick-Villa G, Ko A, O’Connell TX, Guenther JM, Danial T, DiFronzo LA. Does tumor burden limit the accuracy of lymphatic mapping and sentinel lymph node mapping? Cancer J 2002;8:445-50. 12. Kitagawa Y, Watanabe M, Hasegawa H, Yamamoto S, Fujii H, Yamamoto K, et al. Sentinel node mapping for colorectal cancer with radioactive tracer. Dis Colon Rectum 2002; 45:1476-80. 13. Meyer JS. Sentinel lymph node biopsy: strategies for pathologic examination of the specimen. J Surg Oncol 2001; 69:212-8.

Bembenek et al 505

Surgery Volume 135, Number 5

14. Schwartz GF, Guiliano AE, Veronesi U. Proceedings of the consensus conference on the role of sentinel lymph node biopsy in carcinoma of the breast April 19 to 22, 2001, Philadelphia, Pennsylvania. Cancer 2002;94:2542-51. 15. Bilchik AJ, Saha S, Tsioulias GJ, Wood TF, Morton DL. Aberrant drainage and missed micrometastases: the value of lymphatic mapping and focused analysis of sentinel lymph nodes in gastrointestinal neoplasms. Ann Surg Oncol 2001; 8:82-5. 16. Feig BW, Curley S, Lucci A, Hunt KK, Vauthey JN, Mansfield PF, et al. A caution regarding lymphatic mapping in patients with colon cancer. Am J Surg 2001;182:707-12. 17. Bilchik AJ, Nora D, Tollenaar RA, van de Velde CJ, Wood T, Turner R, et al. Ultrastaging of early colon cancer using lymphatic mapping and molecular analysis. Eur J Cancer 2002;18:977-85. 18. Bilchik AJ, Saha S, Wiese D, Stonecypher JA, Wood TF, Sostrin S, et al. Molecular staging of early colon cancer on

19.

20.

21.

22.

23.

the basis of sentinel node analysis: a multicenter trial. J Clin Oncol 2001;19:1128-36. Fitzgerald TL, Khalifa MA, Al Zahrani M, Law CHL, Smith AJ. Ex vivo sentinel lymph node mapping in colorectal cancer: a feasibility study. J Surg Oncol 2002;80:27-32. Wong JH, Steineman S, Caldena C, Bowles J, Namiki T. Ex vivo sentinel node mapping in carcinoma of the colon and rectum. Ann Surg 2001;233:512-21. Merrie AEH, van Rij AM, Phillips LV, Rossaak JI, Yun K, McCall JL. Diagnostic use of the sentinel node in colon cancer. Dis Colon Rectum 2001;44:410-7. Paramo JC, Summerall J, Poppiti R, Mesko TW. Validation of sentinel node mapping in patients with colon cancer. Ann Surg Oncol 2002;9:550-4. Trocha SD, Nora DT, Saha SS, Morton DL, Wiese D, Bilchik AJ. Combination probe and dye-directed lymphatic mapping detects micrometastases in early colorectal cancer. J Gastrointest Surg 2003;7:340-6.