The Template of the Primary Lymphatic Landing Sites of the Prostate Should Be Revisited: Results of a Multimodality Mapping Study

Transcription

1 european urology 53 (2008) available at journal homepage: Prostate Cancer The Template of the Primary Lymphatic Landing Sites of the Prostate Should Be Revisited: Results of a Multimodality Mapping Study Agostino Mattei a, Frank G. Fuechsel b, Nivedita Bhatta Dhar a, Sebastian H. Warncke a, George N. Thalmann a, Thomas Krause b, Urs E. Studer a, * a Department of Urology, University Hospital of Bern, Switzerland b Division of Nuclear Medicine, University Hospital of Bern, Switzerland Article info Article history: Accepted July 26, 2007 Published online ahead of print on August 3, 2007 Keywords: Fusion imaging Lymphadenectomy Prostate cancer SPECT Abstract Objectives: To map the primary prostatic lymphatic landing sites using a multimodality technique. Methods: Thirty-four patients with organ-confined prostate cancer (ct1 ct2; cn0) underwent single-photon emission computed tomography fused with data from computed tomography (SPECT/CT) (n = 33) or magnetic resonance imaging (SPECT/MRI) (n = 1) 1 h after ultrasoundguided intraprostatic injection of technecium (Tc-99m) nanocolloid. The presence of lymph nodes (LNs) containing Tc-99m was confirmed intraoperatively with a gamma probe. A backup extended pelvic lymphadenectomy (PLND) was performed to preclude missed primary lymphatic landing sites. The SPECT/CT/MRI data sets were used to generate a threedimensional projection of each LN site. Results: A total of 317 LNs (median, 10 per patient; range, 3 19) were detected by SPECT/CT/MRI, 314 of which were confirmed by gamma probe. With an extended PLND, two thirds of all primary prostatic lymphatic landing sites are resected compared with only one third with a limited PLND. Conclusions: The multimodality technique presented here enables precise mapping of the primary prostatic lymphatic landing sites. PLND for prostate cancer should include not only the external and obturator regions as well as the portions medial and lateral to the internal iliac vessels, but also the common iliac LNs at least up to the ureteric crossing, thus removing approximately 75% of all nodes potentially harbouring metastasis. # 2007 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Corresponding author. Department of Urology, University Hospital of Bern, CH-3010 Bern, Switzerland. Tel ; Fax: address: urs.studer@insel.ch (U.E. Studer) /$ see back matter # 2007 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi: /j.eururo

2 european urology 53 (2008) Introduction Prostate cancer disseminates initially and primarily to regional LNs [1]. The histological status of pelvic LNs remains the most significant prognostic factor in prostate cancer [2]. Therefore, precise anatomical knowledge of prostate lymphatic drainage is crucial when considering the sites of LN metastases as well as defining the extent of lymphadenectomy or volume for high-dose conformal radiation. Evaluation of the quality of lymphadenectomy, as measured by the number of LNs removed, is hampered by the lack of studies investigating the primary lymphatic landing sites. Whilst the optimal field of lymphadenectomy is still debated, evidence exists that extended PLND enhances the accuracy of surgical staging and has potential therapeutic benefits in patients with not only positive but also negative nodes [3 5]. Similarly, limited information on prostate lymphatic drainage has been an obstacle for planning intensity-modulated radiation therapy. The pathways of prostate lymphatics were described more than 100 yr ago, without however detailing the percentage distribution and categorisation of either primary or secondary lymphatic landing sites [6]. Several studies have since suggested areas in which positive nodes may be found but do not locate the nodes anatomically within each region [7,8].Their location is thus inherently subjective, allowing surgeons to assign the same node to different locations. In the case of a node in the bifurcation of the common iliac artery, for example, one institution may call it external iliac, another internal iliac, and yet another common iliac. Greater clarity regarding the lymphatic landing sites may be achieved by intraprostatic injection of Tc-99m nanocolloid combined with intraoperative use of a gamma probe. However, the limitations of this technique must be clearly understood. Planar lymphoscintigraphy for preoperative mapping is hampered by its poor sensitivity and inability to anatomically identify all nodes. Limited detection with underestimation of the actual number of nodes has been highlighted [9]. Intraoperative gamma probe detection of nodes requires direct contact, making it unlikely that all nodes will be identified and impossible to know whether all have been identified. Furthermore, sentinel scintigraphy is not an approach suited for identifying all LN metastases. In patients with a serum prostate-specific antigen (PSA) above 20 ng/ml and significant nodal disease, lymphatic drainage can be blocked by cancer cells, preventing technecium from accumulating in a significant number (31%) of positive nodes, which consequently go undetected [10]. Single-photon emission computed tomography (SPECT) fused with computed tomography (SPECT/ CT) or magnetic resonance imaging (SPECT/MRI) data has improved spatial resolution and orientation, thus allowing more precise localisation of Tc-99m containing nodes. To determine the prostate s primary lymphatic landing sites, we applied SPECT/CT/ MRI after intraprostatic injection of Tc-99m nanocolloid, which was verified with intraoperative use of a gamma probe and controlled by a systematic backup PLND. To avoid false-negative nodes, only patients without histological evidence of LN metastases were analysed. Using this combined approach, we created a composite map of the prostate s primary lymphatic landing sites and studied the impact of such information on PLND or radiation therapy of prostate cancer patients. 2. Methods 2.1. Patients Thirty-four patients with a median age of 63 yr (range, 51 72), a median serum PSA of 8 ng/ml (range, ), biopsy-confirmed carcinoma of the prostate T1 to T2, and no metastatic disease (cn0, cm0) based on physical examination and CT or MRI of the pelvis were prospectively enrolled. Exclusion criteria were LNs greater than 1 cm in diameter, previous pelvic radiotherapy, androgen ablation therapy, and a history of pelvic surgery. Patients with metastatic adenopathy found on histopathology were also excluded. All 34 patients gave written informed consent SPECT technique One day prior to surgery, 2 aliquots of 100 MBq in 0.6 ml of Tc-99m nanocolloid (Nanocoll 1 ; GE Healthcare, Amersham Health, NJ, USA) were injected into each lobe of the prostate with a Chiba needle ( mm). Under ultrasound guidance, fractions were injected bilaterally, one each into the peripheral and central zone of the prostatic apex, midportion, and base, with care to obtain homogenous distribution. All injections were performed or supervised by the same urologist and nuclear medicine physician. Care was taken to avoid spillage of the radiocolloid into the rectum. To reduce possible contamination, all patients used disposable underwear and were asked to empty their bladder and bowel after the injection. One hour after injection of Tc-99m nanocolloid, SPECT of the pelvis and abdomen (gamma camera, IRIX; Philips, Milpitas, CA, USA) was performed according to a standardised protocol (LEGAP collimator, matrix, 38/45 s per step, with patients in the supine position) SPECT/CT/MRI fusion and interpretation All patients underwent preoperative CT (n = 33) or MRI (n = 1, owing to renal impairment) of the pelvis for initial diagnostic

3 120 european urology 53 (2008) purposes. Data from all examinations were available in digital form. Standardised protocols were used for both threedimensional (3D) SPECT and image postprocessing (count truncation of injection site maxima, iterative reconstruction, four iterations, Butterworth postfiltering, reconstructed slice thickness 3.3 mm). Reconstructed coronal, sagittal, and transverse data sets were overlain with the corresponding CT or MRI images with the use of commercially available software (esoft#, Syngo 3D Fusion TM ; Siemens, Erlangen, Germany). SPECT/CT/MRI fusion imaging was done by the same specialist in all cases. Each node was depicted in transverse, sagittal, and coronal views. Images were read by two examiners with regards to number of nodes and anatomical localisation LN identification at surgery To decrease background radioactivity, all patients were catheterised transurethrally prior to surgery. Surgery was performed by three experienced surgeons. After midline incision and mobilisation of the peritoneum, prostate lymphatic drainage regions I VI (Fig. 1) were systematically explored for LNs with the use of a standard gamma probe (C-Track#, Care Wise Medical Products Corporation; AEA Technology, Morgan Hill, CA, USA) with a wide-aperture collimator (for greater sensitivity) wrapped in a sterile polyvinyl chloride sheath. The injection site (prostate) was shielded with a malleable sterilised lead plate to decrease superfluous radioactivity and to improve the signal-to-noise ratio for LN detection. Background radioactivity was determined over the sacrum and/or vena cava. A focal signal intensity of twice or more the background signal was defined as indicative of LNs. The search was performed by systematic screening whilst keeping the gamma probe very close to the vascular structures at different angulations along all areas with suspected LNs from the deep pararectal/presacral regions to the external iliac vessels and up to the inferior mesenteric artery. A nuclear medicine specialist was present in the operating room and checked for completeness of the surgical localisation by comparing the number and localisation of the LNs reported on SPECT/CT/MRI fusion imaging with those found at surgery. Those LNs with the highest intraoperative count rates were resected first to reduce interference and to improve discrimination of the remaining nodes with lower count rates. Intraoperatively located LNs that could not be resected for technical reasons or because the necessary enlargement of the operation field would place the patient at a risk disproportionate to the potential gain from resection were counted as identified LNs. LNs in para-aortic and paracaval locations were excised only if an extraperitoneal approach was possible. To detect any radioactivity missed on the first scan, we scanned all regions again after LN resection. Gamma probe guided LN dissection was always followed by bilateral extended PLND along the major pelvic vessels including the internal iliac, external iliac, and obturator regions. To identify any previously missed LNs, we screened the removed lymphatic tissue ex vivo for radioactivity with a gamma camera optimised for sensitivity (no collimator, planar images, matrix, full-field mode acquisition over 10 min). The specimens were sent separately according to anatomical region for histological analysis. The 3D localisation of each LN was transferred to a projection model of the pelvis on the basis of SPECT/CT/MRI fusion imaging. The transference was done particularly with respect to the location of the individual node in relation to the pelvic vascular structures and ureters (Fig. 2a and b). Fig. 1 Regions of pelvic lymph node dissection.

4 european urology 53 (2008) Fig. 2 Primary lymphatic landing sites of the prostate transferred from 3D datasets and superimposed into idealized anterior-posterior and medio-lateral projection of the pelvis (n = 317 for 34 patients). All LNs depicted 3D reconstruction of single-photon emission computed tomography/computed tomography/magnetic resonance imaging (SPECT/CT/MRI) data sets and confirmed by surgery were subsequently transferred into a template projection of the pelvis. LNs are color coded in relation to the vascular structures and ureters. Superimposed are the boundaries of pelvic lymphadenectomy (PLND) subdivided into different regions. (a) Primary lymphatic landing sites of the prostate in anterior-posterior projection of the pelvis. (b) Primary lymphatic landing sites of the prostate in medio-lateral projections of the pelvis. Figs. (b)a. and (b)b. are reconstructions of the left and right hemipelvis, respectively, in the median view and therefore only the LNs of the left and right side are respectively shown. To prevent complications from the PLND, we ligated all lymphatics from the lower extremities, placed a drain on each side of the pelvis, and injected low-molecular heparin into the upper arm. 3. Results Bilateral lymphatic drainage was seen in all patients on scintigraphy. Of the 317 LNs detected by SPECT/CT/MRI fusion imaging (median, 10; range, 3 19), intraoperative gamma probe found 314 (99%). Of the 314 LNs detected intraoperatively by gamma probe, 300 could be surgically removed. Fourteen LNs were not removed because of difficult surgical access or increased risk of complications in the deep pararectal or para-aortic/retroaortic or caval areas. Histological examination confirmed all LNs to be free of metastases. The 317 LNs found on SPECT/CT/MRI fusion imaging were distributed as follows: external iliac and obturator fossa (n = 120), internal iliac (n = 81), presacral and pararectal (n = 26), common iliac (n = 50), para-aortic/paracaval (n = 38), and inguinal (n =2) (Fig. 3). Only 38% of the LNs were located within the area of the commonly performed limited PLND area, dorsal to and along the external iliac vein

5 122 european urology 53 (2008) Fig. 3 Anatomical distribution of 317 lymph nodes of the prostate of all 34 patients based on single-photon emission computed tomography/computed tomography/magnetic resonance imaging (SPECT/CT/MRI) fusion imaging confirmed by surgery. and along the obturator nerve. Only 63% were located in the region of extended PLND, which also includes the LNs both medial and lateral to the internal iliac vessels. On the basis of these findings we developed a new bilateral extended PLND template for prostate cancer (Fig. 4). With the bilateral extended backup PLND along the internal iliac, external iliac, and obturator regions, another 569 LNs not containing Tc-99m nanocolloid (median, 15; range, 6 35) were removed. Histological examination thus revealed a total of 869 LNs (median, 26; range, 13 44). One patient required surgical exploration 1 d after resection of LNs in the para-aortic region at the level of the second lumbar vertebral body because of retroperitoneal bleeding. Following this complication in the 18th patient of our series, nodes in the para-aortic/paracaval region were excised only if mobilisation of the peritoneum permitted sufficient exposure and access to the surgical field. No further complications were encountered. 4. Discussion The adverse prognosis for metastatic disease to the pelvic LNs in prostate cancer patients is universally Fig. 4 Anatomical extent of classical extended pelvic lymphadenectomy (PLND) (a) and of a proposed (new) extended PLND (b) for prostate cancer. (a) Area of classical extended template PLND for prostate cancer encompassing the nodes along the major pelvic vessels including the internal iliac, external iliac and obturator regions to the iliac bifurcation (yellow and orange areas). (b) Area of the proposed (new) extended template PLND extending along the common iliac vessels to the ureteric crossing (pale red area).

6 european urology 53 (2008) accepted. Sometime during the process of metastatic spread, a therapeutic window exists for treatment [11]. Knowledge of the landing sites of prostate LN metastases would be helpful for management of patients with high-risk prostate cancer. Wawroschek et al [12] used intraprostatic injection of Tc-99m nanocolloid combined with intraoperative gamma probe to address limitations of previous LN mapping studies, yet this technique also fails to be complete [12]. Nodes outside the fields where comprehensive search is performed are inadvertently missed. Systematic screening requires maintaining the gamma probe close to vascular structures at various angulations along all areas containing possible LNs, ranging from the pararectal/presacral region to the inferior mesenteric artery. Furthermore, intraoperative gamma probe guided LN detection is time consuming if all nodes are to be evaluated and removed; the surgical time is more than twice that of an extended PLND [13]. To elucidate the pattern of prostatic lymphatic spread, we applied SPECT/CT/MRI after intraprostatic injection of Tc-99m nanocolloid, which was verified with intraoperative use of a gamma probe and controlled by a backup extended PLND. This combined technique revealed a more complex lymphatic drainage than was previously appreciated. It demonstrated prostate lymphatic drainage to the pelvic LNs as well as those in the para-aortic/ paracaval region. Only two thirds of all nodes in the pelvis are located along the external and internal iliac vessels and in the obturator fossa. Another 16% are found along the common iliac vessels, and 8% in the presacral/pararectal regions. Although the principal location of the nodes is the pelvis, a considerable percentage (12%) are situated along the aorta and vena cava as high as the origin of the inferior mesenteric artery. Interestingly, more than twice as many nodes (n = 27) are located along the aorta as along the vena cava (n = 11). This asymmetric distribution can conceivably be attributed to an embryological symmetric growth of the venae cardinales caudales (accompanied by lymphatics) followed by asymmetric growth of the vena cava. The Takashima [14] group defined completeness of an LN resection by dividing the number of nodes detected in vivo by the number of nodes detected in vivo and ex vivo with the scintillation camera after extended PLND encompassing the common, external, and internal iliac vessels; obturator fossa; and presacral regions. Even this procedure seems to be incomplete on the basis of our results because it omits nodes outside the lymphadenectomy template. Indeed, our 34 preoperative SPECT/CT/MRI fusion imaging data sets depicted 38 nodes in the para-aortic/paracaval region, 26 in the pararectal/ presacral, and 2 in the inguinal region, representing 12%, 8%, and 1% of the total LN count, respectively. Whether the para-aortic/paracaval nodes can be definitely considered primary landing sites or juxtaregional nodes may be questioned. The appearance of these nodes in the SPECT simultaneously with those in the regions close to the prostate 1 h postinjection suggests that they can be defined as primary landing sites. However, until morphological markers are found that allow identification of each lymphatic vessel and the node to which it drains, we cannot be certain. To ensure that no nodes were missed, we included only pn0 patients because nodal metastases can obstruct lymphatic flow, preventing further uptake of Tc-99m nanocolloid in a substantial number of nodes [10]. The clinical implications of the present study are significant. It is to our knowledge the first reported series documenting the prostate s primary lymphatic landing sites. If the location and incidence of primary lymphatic landing sites are known, then the boundaries of lymphadenectomy can be selected and pelvic radiation toxicity reduced by limiting the fields to areas of nodal involvement. Our results demonstrate that the most frequently applied technique of limited PLND (obturator fossa and nodes along the external iliac vein) resects at best 38% of all nodes. The extended variant (obturator fossa, external iliac vessels, lateral and medial regions of the internal iliac vessels) reduces to about one third the percentage of nodes left in situ (Fig. 4). A complete LN dissection up to the origin of the inferior mesenteric artery, however, does not appear justifiable. It would be too demanding with regards to both time and extent of surgical field, with the associated increased potential for complications. Removal of LNs medial to the ureters and the para-aortic/paracaval, pararectal, and presacral regions also increases the risk of injury to the autonomic sympathetic nerves, which may compromise both continence and sexual function [15]. These risks are too high, particularly because it is not known whether resection of extrapelvic nodes up to the inferior mesenteric artery is beneficial to the patient. We therefore suggest that the template of PLND be extended up to only where the ureters cross the common iliac artery if a PLND is indicated (Fig. 4b) [16,17]. Application of our proposed new extended PLND template removes approximately 50% of nodes along the common iliac vessels and thereby removes close to 75% of all prostate primary lymphatic landing sites. It remains open whether

7 124 european urology 53 (2008) the PLND should be extended to the proximal common iliac vessel region and the aortic bifurcation in select patients with a high serum PSA and Gleason score. The availability of 3D conformal radiation treatment and intensity-modulated radiation treatment allows the application of higher radiation doses without increasing the toxicity to normal tissues. Decreasing the target volume by millimeters may significantly reduce treatment-related toxicity [18]. Our results provide radiation oncologists with maps to guide delivery of a target volume of high-dose conformal radiation while minimizing radiation toxicity to nontarget normal tissue. However, the radiation oncologist will also have to balance the risk of missing some nodes against the morbidity of an extended field radiation therapy. Although our combined approach has revealed the primary lymphatic landing sites of the prostate, it is not without limitations. Although more nodes are identified with SPECT/CT/MRI fusion imaging analysis (median, 10 LNs) than reported by most authors using planar imaging (median, 4 6 LNs), or gamma probe (median, 8 LNs) alone, the required acquisition time is longer [13,19]. Furthermore, analysis of SPECT/CT/MRI fusion imaging requires the expertise of a nuclear medicine specialist dedicated to a thorough search for all nodes. 5. Conclusion In summary, our combined modality technique in patients not harbouring LN metastasis seems to be an excellent tool to define the prostate s primary lymphatic landing sites. It reveals that the template of primary lymphatic landing sites is larger than previously appreciated: Nodes can be found up to the inferior mesenteric artery. With a limited PLND, 38% of LNs are at best removed, whereas with an extended PLND approximately 63% are removed. Complete dissection encompassing all regions where primary prostatic lymphatic nodes are detected, however, is hardly feasible because of the increased associated costs, time, extent of surgery, and risk of complications. As a compromise for patients at risk for nodal metastases, we therefore suggest a template encompassing the area covered by classic extended PLND plus the nodes along the common iliac arteries up to where the ureter crosses, thereby removing 75% of all LNs. Likewise the radiation oncologist will have to balance the risk of potentially omitting some nodes against the morbidity of an extended pelvic field. Conflicts of interest The authors have nothing to disclose. References [1] Flocks RH, Culp D, Porto R. Lymphatic spread from prostatic cancer. J Urol 1959;81: [2] Gervasi LA, Mata J, Easley JD, et al. Prognostic significance of lymph nodal metastases in prostate cancer. J Urol 1989;142: [3] Masterson TA, Bianco Jr FJ, Vickers AJ, et al. The association between total and positive lymph node counts, and disease progression in clinically localized prostate cancer. J Urol 2006;175:1320 4, discussion [4] Joslyn SA, Konety BR. Impact of extent of lymphadenectomy on survival after radical prostatectomy for prostate cancer. Urology 2006;68: [5] Daneshmand S, Quek ML, Stein JP, et al. Prognosis of patients with lymph node positive prostate cancer following radical prostatectomy: long-term results. J Urol 2004;172: [6] Cuneo B, Marcille M. Topographie des ganglions-ilio pelviens. Bull Mem Soc Anat Paris ser.jt.iii [7] Heidenreich A, Varga Z, Von Knobloch R. Extended pelvic lymphadenectomy in patients undergoing radical prostatectomy: high incidence of lymph node metastasis. J Urol 2002;167: [8] Bader P, Burkhard FC, Markwalder R, Studer UE. Is a limited lymph node dissection an adequate staging procedure for prostate cancer? J Urol 2002;168: [9] Jansen L, Nieweg OE, Kapteijn AE, et al. Reliability of lymphoscintigraphy in indicating the number of sentinel nodes in melanoma patients. Ann Surg Oncol 2000;7: [10] Weckermann D, Dorn R, Holl G, Wagner T, Harzmann R. Limitations of radioguided surgery in high-risk prostate cancer. Eur Urol 2007;51: (discussion ). [11] Carter HB, Ferrucci L, Kettermann A, et al. Detection of life-threatening prostate cancer with prostate-specific antigen velocity during a window of curability. J Natl Cancer Inst 2006;98: [12] Wawroschek F, Vogt H, Weckermann D, Wagner T, Hamm M, Harzmann R. Radioisotope guided pelvic lymph node dissection for prostate cancer. J Urol 2001;166: [13] Warncke SH, Mattei A, Fuechsel FG, Z brun S, Krause T, Studer UE. Detection rate and operating time required for gamma probe-guided sentinel lymph node resection after injection of technetium-99m nanocolloid into the prostate with and without preoperative imaging. Eur Urol 2007;52: [14] Takashima H, Egawa M, Imao T, Fukuda M, Yokoyama K, Namiki M. Validity of sentinel lymph node concept for patients with prostate cancer. J Urol 2004;171: [15] Burkhard FC, Kessler TM, Fleischmann A, Thalmann GN, Schumacher M, Studer UE. Nerve sparing open radical retropubic prostatectomy does it have an impact on urinary continence? J Urol 2006;176: [16] Schumacher MC, Burkhard FC, Thalmann GN, Fleischmann A, Studer UE. Is pelvic lymph node dissection

8 european urology 53 (2008) necessary in patients with a serum PSA < 10 ng/ml undergoing radical prostatectomy for prostate cancer? Eur Urol 2006;50: [17] Madersbacher S, Hochreiter W, Burkhard F, et al. Radical cystectomy for bladder cancer today a homogeneous series without neoadjuvant therapy. J Clin Oncol 2003;21: [18] Shih HA, Harisinghani M, Zietman AL, Wolfgang JA, Saksena M, Weissleder R. Mapping of nodal disease in locally advanced prostate cancer: rethinking the clinical target volume for pelvic nodal irradiation based on vascular rather than bony anatomy. Int J Radiat Oncol Biol Phys 2005;63: [19] Wawroschek F, Vogt H, Wengenmair H, et al. Prostate lymphoscintigraphy and radio-guided surgery for sentinel lymph node identification in prostate cancer. Technique and results of the first 350 cases. Urol Int 2003;70: