|Home | About | Journals | Submit | Contact Us | Français|
To describe the accuracy of SLN mapping in patients with a preoperative diagnosis of grade 1 endometrial cancer.
A prospective, non-randomized study of women with a preoperative diagnosis of endometrial cancer and clinical stage I disease was conducted. A subset analysis of patients with a preoperative diagnosis of grade 1 endometrial endometrioid cancer was performed. All patients had preoperative lymphoscintigraphy with Tc99m on the day of or day before surgery followed by an intraoperative injection of 2cc of isosulfan or methylene blue dye deep into the cervix or both cervix and fundus. All patients underwent hysterectomy, bilateral salpingo-oophorectomy, and regional nodal dissection. Hot and/or blue nodes were labeled as SLNs and sent for histopathological analysis.
Forty-two patients with a preoperative diagnosis of grade 1 endometrial carcinoma treated from 3/06–8/08 were identified. Twenty-five(60%) had laparoscopic surgery; 17(40%) were treated by laparotomy. Preoperative lymphoscintigraphy visualized SLNs in 30 patients(71%); intraoperative localization of the SLN was possible in 36(86%). A median of 3 SLNs(range, 1–14) and 14.5 non-SLNs(range, 4–55) were examined. In all, 4/36(11%) had positive SLNs—3 seen on H&E and 1 as cytokeratin-positive cells on IHC. All node-positive cases were picked up by the SLN; there were no false-negative cases. The sensitivity of the SLN procedure in the 36 patients who had an SLN identified was 100%.
Sentinel lymph node mapping using a cervical injection with combined Tc and blue dye is feasible and accurate in patients with grade 1 endometrial cancer and may be a reasonable option for this select group of patients. Regional lymphadenectomy remains the gold standard in many practices, particularly for the approximately 15% of cases with failed SLN mapping.
Comprehensive surgical staging for endometrial carcinoma is accepted by most gynecologic oncologists as the standard of care for the majority of patients with grade 2–3 endometrioid endometrial cancers as well as patients with clear cell and papillary serous tumors. However, these recommendations and the surgical treatment of patients with a preoperative diagnosis of grade 1 endometrioid endometrial cancer remains an area of significant debate, with a wide variation in the intraoperative management of the retroperitoneal nodes. Some patients are not surgically staged, some are comprehensively staged, and some are limitedly staged depending on the surgeon’s intraoperative assessment of the tumor’s risk for metastasis. The utility and value of intraoperative uterine examination by the surgeon or by a pathologist at frozen-section evaluation of depth of invasion and tumor size also remains variable among different institutions and practitioners; specifically, intraoperative frozen-section analysis for histologic grade and depth of myometrial invasion may not correlate well with final pathologic grade and stage .
Two significant contributors to the inconsistency in intraoperative management and universal acceptance of comprehensive surgical staging are the rarity of nodal metastases in true grade 1 endometrioid adenocarcinomas on final uterine pathology  and the excellent overall prognosis of women with a final diagnosis of grade 1 endometrial cancer . In addition, there may be concerns about overtreating a potentially low-risk group of women who may have a prolonged survival, particularly when a comprehensive regional lymphadenectomy is performed, which may have long-term effects, including symptomatic leg lymphedema .
However, certain pathologic and clinical characteristics of women with a preoperative diagnosis of grade 1 endometrial cancer need further clarification. It is well documented that not all patients who present with a preoperative diagnosis of grade 1 endometrial carcinoma have a final diagnosis of grade 1 endometrial cancer; therefore, if not comprehensively staged, nearly 15% of patients deemed to have grade 1 lesions preoperatively will actually have higher grade lesions on final post-hysterectomy pathology. These patients would have otherwise been eligible and benefited from comprehensive staging . In addition, identification of nodal metastasis, although rare, will likely have profound effects on postoperative management and adjuvant therapy. Therefore, our challenge is to identify a surgical technique that provides accurate staging information about nodal status while avoiding the potential for overtreating low-risk patients and undertreating patients with metastatic disease; in other words, identifying a more accurate surgical technique to stage endometrial cancer and avoid unnecessary morbidity.
The objective of this manuscript was to describe our initial experience with sentinel lymph node (SLN) injections and mapping in women with a preoperative diagnosis of grade 1 endometrial cancer and to determine the sensitivity of this technique in an effort to suggest a new treatment algorithm for women presenting with grade 1 disease on endometrial biopsy or dilatation and curettage.
A prospective, non-randomized study of women presenting with a preoperative diagnosis of endometrial cancer with clinical stage I disease was conducted. This study reports on the subset of patients with preoperative grade 1 endometrioid adenocarcinoma. Other inclusion criteria included the following: a performance status of 0, 1, 2, or 3 by the Gynecologic Oncology Group criteria; and patients agreed to undergo total hysterectomy, removal of both adnexae, and bilateral regional lymphadenectomy via laparotomy or laparoscopy. For the purposes of this report, we defined any blue or hot node that was identified and confirmed by pathology in either the right or left pelvic or paraaortic retroperitoneum as a detected SLN. The study was approved by our Institutional Review Board and all patients signed an approved informed consent. Standard statistical methods were used when appropriate.
Preoperative lymphoscintigraphy was obtained following injection of 0.1–0.5 mci radiolabeled filtered (0.22 millipore filter) Tc99 microsulfur colloid in 0.1–0.5ml volume. This was injected using a 27-gauge Potocky® needle (Cooper Surgical, Trumbull, CT) into the stroma of the cervix at the 3 and 9 o’clock positions by a member of the gynecology service with the assistance of a member from the nuclear medicine department. The injection was prepared by the nuclear medicine department, which also disposed of the syringe and other materials. The injection of radioisotope and lymphoscintigram were performed either the morning of or day prior to surgery, depending on the attending’s and patient’s preference. A preoperative lymphoscintigram was obtained after the injection, and 2 series of pictures were taken— immediate “dynamic images” and subsequent “static images” to localize the nodes. Patients spent approximately 1 hour in the nuclear medicine department.
The injection of blue dye into the cervix was performed in the operating room at the time of the examination under anesthesia. A spinal needle was used to inject a total of 2 cc of isosulfan (Lymphazurin® 1%, Tyco Healthcare, Norwak, CT) or methylene blue dye (Methylene Blue 1%, American Regent, Inc. Shirley, NY) into the cervical stroma at the 3 and 9 o’clock positions (Figure 1). Each injection was 1 cc. After the cervical injection of blue dye, the patient was prepped and draped in the usual sterile fashion. The procedure was continued as planned either through laparoscopy or laparotomy. For patients enrolled in the second half of the study cohort, 2 additional blue dye injections into the fundus were added—1 cc in the anterior mid fundus and 1 cc in the posterior mid fundus using a 22-gauge 36-cm Nezhat-Dorsey aspiration/injection laparoscopic needle (Davol, Inc. Cranston, RI) (Figure 2). The SLN identification and removal was performed first after obtaining the pelvic washings. The completion lymphadenectomy was performed either prior to or after the hysterectomy.
A handheld gamma probe was used to record counts and detect “hot” nodes, defined as nodes with counts detected after background quenching (for laparoscopic cases we used a 10-mm C-Trak® probe by Care Wise Medical Products, Morgan Hill, CA). Blue nodes were detected with the naked eye. The SLN(s) that were identified were labeled as “hot” and/or “ blue”. The anatomic location of the node(s) and counts (ex vivo) were recorded on a standardized data collection sheet. In addition, a post-excision bed count of each lymphatic basin was recorded after removal of the SLN(s) in each basin. The SLNs were not sent for a frozen-section analysis on a routine basis. Following the identification of the SLN(s), a bilateral pelvic and paraaortic node dissection was performed. The paraaortic node dissection was limited to the vicinity of insertion of the right ovarian vein into the vena cava on the right side and the vicinity of the inferior mesenteric artery on the left side. Lymph nodes from the lymphadenectomy were anatomically labeled as right or left: external iliac, internal iliac, obturator, common iliac, and paraaortic. It should be noted that the proximal obturator nodes and the internal iliac nodes are anatomically difficult to distinguish and frequently overlap. The assignment of SLN to internal iliac versus obturator was left up to the surgeon removing the SLN, but in general, proximal obturator nodes overlying the internal iliac vessels were considered internal iliac and the more distal obturator nodes ventral to or surrounding the obturator nerve were considered obturator.
The pathologic protocol for SLN evaluation at our institution is as follows: from each paraffin block lacking metastatic carcinoma appreciable in a routine section stained with hematoxylin and eosin (H&E), 2 adjacent 5-micron sections were cut at each of two levels 50 microns apart. At each level, one slide was stained with H&E and the other with immunohistochemistry (IHC) using the anti-cytokeratin AE1:AE3 (Ventana Medical Systems, Inc., Tucson, AZ) for a total of 4 slides per block.
Forty-two patients with a preoperative diagnosis of grade 1 endometrial endometrioid carcinoma treated between 3/06 and 8/08 were enrolled. Patients were diagnosed based on an office endometrial biopsy in 17 cases (40%) or dilatation and curettage in 25 cases (60%). Median age was 60.5 years (range, 34–82). Median body mass index (BMI) was 29.3 kg/m2 (range, 19–61 kg/m2). Twenty-five cases (60%) were performed by laparoscopic surgery, and 17 (40%) were treated by laparotomy. The selection of the surgical approach was based on the attending physician’s discretion.
Injection sites were specified by the study protocol. Twenty-one patients (50%) had cervical injections (Tc99m and 2 cc blue dye at the 3 and 9 o’clock positions), and the other half of the cohort had cervical injections with Tc99m and blue dye, as the first group did, but also had 2 fundal injections with 1 cc blue dye each in the mid anterior and posterior fundus. The type of blue dye used was based on availability from the pharmacy during the study period—isosulfan blue in 35 cases (83%) and methylene blue in 7 cases (17%).
Preoperative lymphoscintigraphy with cervical injection of Tc99m was performed the day before surgery in 28 patients (67%) and the same day in 14 patients (33%). Preoperative lymphoscintigraphy visualized SLN in 30 (71%) of 42 patients; moreover, intraoperative localization of the SLN was possible in a total of 36 patients (86%). Three (25%) of the 12 cases with negative preoperative Tc99m lymphoscintigraphy were mapped successfully with blue dye in the operating room.
A total of 145 SLNs were identified and examined by pathologists. Of these, 58 SLNs (40%) were located in the right pelvis and 55 (38%) were located in the left pelvis; in addition, 4 (3%) were in right paraaortic nodes and 1 (0.6%) was in left paraaortic nodes (Table 1). Figure 3 summarizes the distribution of identified SLNs and percentage of SLNs at each site. Overall, the majority 129 (89%) of the SLNs in grade 1 endometrial cancer were located in pelvic lymph nodes (external/internal iliac or obturator nodes) commonly included in the templates of routine bilateral total pelvic lymphadenectomy (Figures 4 and and5).5). The most common anatomic sites where SLNs were identified, in decreasing order, were as follows: internal iliac nodes, 52 (36%); external iliac, 43 (30%); obturator, 34 (23%); common iliac, 11 (8%) (Figure 6); and paraaortic, 5 (3%) (Figure 7).
A median of 3 SLNs (range, 1–14) and 14.5 non-SLNs (range, 4–55) were examined. In all, 4 (11%) of the 36 patients who had successful mapping had positive SLNs; in 3 cases, these positive SLNs were seen on H&E, and 1 case was detected as cytokeratin-positive cells on IHC. All node-positive cases were picked up by SLN mapping, and there were no false-negative cases. None of the 6 cases that were not mapped had positive nodes on completion lymphadenectomy. Overall, the sensitivity of the SLN procedure in the 36 grade 1 patients who had an SLN identified was 100%.
All patients had a preoperative diagnosis of grade 1 endometrial cancer. This was confirmed on final post-hysterectomy analysis in 32 cases (76%), 9 cases (21%) were upgraded to grade 2, and 1 case (2%) to grade 3. Final pathologic International Federation of Gynecology and Obstetrics (FIGO) stage was as follows: IA, 24 (57%); IB, 9 (21%); IC, 1 (2%); IIA, 1 (2%); IIIA, 3 (7%); IIIC, 3 (7%); and IVB, 1 (2%). Lymph-vascular invasion was identified on final pathology in 7 cases (17%). The median uterine weight was 106 g (range, 33–617g).
The 4 SLN-positive cases were upstaged to stage IIIC in 3 cases (final grade remained well-differentiated in 2 cases and upgraded to moderately differentiated in 1 case) and to stage IVB in 1 case due to intra-abdominal metastatic nodules confirmed as grade 3 on final pathology.
Of the 32 cases with a confirmed final post-hysterectomy diagnosis of grade 1 endometrial cancer, 2 (6%) had positive pelvic nodes (stage IIIC). Final pathology revealed a myometrial invasion of 7/19 mm with lymphovascular invasion and 2 out of 18 positive nodes in one case and 17/18 mm with lymphovascular invasion and 3 out of 14 positive lymph nodes in the second case. Of the 4 cases with positive SLNs, the final results of the non- SLNs removed were as follows (reported as number of positive nodes/total number of non-SLNs examined): 0/13, 1/12, 0/16, and 0/16. In other words, the SLN was the only positive node in three-fourths of cases with nodal metastasis; and in 1 case, additional non- SLNs were also positive.
Six patients (14%) had a failed mapping, with no SLN identified. The median age was 64.5 years (range, 51–82 years), the median BMI was 33.3 kg/m2 (range, 21–61 kg/m2), the median uterine weight was 84 g (range, 54–220 g), and the median number of non-SLNs examined was 17 (range, 9–26). The final grade was confirmed as grade 1 in 5 of 6 cases, and 1 case was upgraded to grade 2. No patient had lymph-vascular invasion or positive lymph nodes, and final FIGO stage included stage IA (3) and stage IB (3). It is unclear why mapping failed in these 6 cases, but the following explanation may have contributed to this technical failure. Two patients were morbidly obese, one with a BMI of 61 (11 negative non-SLNs) and the other with a BMI of 40 (18 negative non-SLNs), which may have contributed to the difficulty in visualizing SLNs; moreover, 3 failed cases were performed by surgeons with beginner’s experience (first 1–2 cases) in SLN mapping early on in the study, and the last failed case was an 82-year-old female (BMI=33) with very tortuous iliac vessels that we were unable to identify SLN by either Tc99m or blue dye; she had stage IA, grade 1 disease, with 9 negative non-SLNs.
The addition of 2 fundal injections in the latter part of the study did not appear to improve on the SLN detection rates, with 4 of 21 failed mappings seen with the cervical injection group (the earlier group) compared to 2 of 21 with combined cervical and fundal injections (P=0.4); the trend toward improvement in detection rates may be more related to increased experience with SLN mapping techniques as the study progressed rather than the addition of the fundal injections; in addition, the type of blue dye used did not influence our detection rates, with 4 of 35 failed mappings in the isosulfan group compared to 2 of 7 failed mappings in the methylene blue group (P=0.3).
The search continues for a surgical technique that provides accurate pathologic information about the nodal status of patients with grade 1 endometrial cancer without overtreating potentially low-risk patients and undertreating patients with metastatic disease. SLN mapping in women with a preoperative diagnosis of grade 1 endometrial cancer may be an acceptable solution. This is usually a low-risk group for nodal metastasis; however, missing nodal disease in these select cases will likely have a detrimental oncologic outcome.
The use of SLN mapping is now well established in the treatment of melanoma  and breast cancer . The SLN hypothesis proposes the SLN(s) as the first node(s) in a regional lymphatic basin that receive lymph flow from the primary tumor. The histologic status of the SLN may accurately predict the status of the regional lymphatic basin. Thus, if the SLN status is negative, a regional lymphadenectomy may be avoided. By avoiding a systematic lymphadenectomy, the morbidity associated with this procedure can be reduced. The morbidity associated with a complete pelvic lymphadenectomy, such as leg lymphedema, infected or symptomatic pelvic lymphocysts, and chylous ascites in rare cases, may be reduced. The success of this technique in breast and melanoma has led to its application in the field of gynecologic oncology. SLN mapping for uterine corpus malignancies, however, proved to be a significant challenge to gynecologic oncologist interested in the utilization of SLN mapping techniques .
In the current study, the majority of the SLNs (129 [89%]) were located in pelvic lymph nodes commonly included in the templates of routine pelvic lymphadenectomy (Figure 4); moreover, it is important to mention that these SLNs were usually located on the medial and ventral aspect of the iliac vessels, and none were in the deep circumflex iliac area, confirming previous observations that the circumflex iliac nodes are usually not sentinel to uterine drainage .
The rate of detection of SLNs is related to anatomic and technical factors, such as the adequacy of the injection, the site of injection, the substance injected, and the techniques used to identify the SLN. Undoubtedly, a successful SLN mapping requires a dedicated surgical team as well as a dedicated radiology and gynecologic pathology team. Similar to any other surgical procedure, there is a learning curve, and with more cases performed the detection rates will likely improve. There are 3 injection sites currently described in lymphatic mapping of endometrial cancer: 1) the uterine corpus subserosal/myometrial, 2) the cervix, and 3) the endometrium via hysteroscopy.
Several studies have used the subserosal myometrium as the site of injection of the marker to identify the SLN (Table 2) [10–16]. Burke et al.  published the pioneer report on the identification of SLNs in endometrial cancer. In their series of 15 patients, at the time of laparotomy, they occluded the tubes with hemoclips and injected isosulfan blue into the subserosal myometrium of the uterine fundus at 3 midline sites. Identification and removal of dye-containing nodes was performed followed by the standard staging procedure. Deposition of dye in at least 1 node was observed in 67% of cases. Two out of 4 metastatic nodes did not contain blue dye (false-negative rate of 50%). Holub et al.  and Gien et al.  reported similar detection rates of 61.5% and 56%, respectively, using a similar technique. On the other hand, Echt et al. failed to identify any SLN in their series of 8 patients .
Li et al.  reported a detection rate of 75% using 1% Methylene blue injected into the fundal subserosal myometrium and at the uterine isthmus. In their series of 20 patients, after the initial 5 patients in whom no SLN was identified due to technical details, the authors identified an SLN in the remaining 15 patients. Therefore, the 75% detection rate reported may be an underestimation of the real detection rate with this technique. Altgassen et al.  recently reported on a new approach for identification of the SLN during laparotomy. They used 8 subserosal injection sites (4 ventrally and 4 dorsally) in their series, with a 92% detection rate, which is the highest detection rate described for subserosal injection. This technique may be similar to the technique described by Li et al.  and seems to indicate that the detection rate increases with the number of injections at different sites of the uterine corpus.
Frumovitz et al.  recently reported on their series of 18 patients with clinical stage I endometrial cancer. This is the only study using both radioactive colloid and blue dye injection in the subserosal myometrium. Identification of SLNs was performed by a handheld gamma counter or by direct observation of blue discoloration of lymph nodes. The detection rate for an SLN was only 45%. Surprisingly, this technique was less effective when only blue dye was used by the same investigators . The reason for this unexpected low rate of detection is not clear.
Undoubtedly, this is the easiest and most convenient injection site for both blue dye and Tc99m. Several studies have used the cervix as the site of injection, either exclusively [17–21] or in combination with an injection into the subserosal myometrium [12,22] (Table 3) [12,17–22]. The rates of identification of an SLN varied from 80% to 100%. In the studies using both blue dye and radioactive colloid, the detection rate was exactly the same for both the radioactive tracer and the blue dye in 3 studies [17–19], while a higher detection rate was achieved by the radioactive tracer in the other 2 studies [20,21]. In the latter two studies, however [20,21], 3 patients had metastatic nodes that were stained blue but had no radioactivity detected, leading to the consideration not to abandon the blue dye as an adjunct to the radioactive tracer if this technique is being used. Holub et al. used a combination of cervical and subserosal injections of blue dye [12,22] and also reported a detection rate in the 80% range. A theoretical concern with cervical injection is that the nodal spread patterns are somewhat different for cervical and endometrial cancers, so it is debatable whether injections into the cervix could give reliable information on SLNs in endometrial cancer. And this was a reason why half of our cohort also was given 2 fundal injections with blue dye. From this study, however, we demonstrated that deep cervical injections at the 3 and 9 o’clock positions (corresponding to the paracervical and parametrial lymphatics) with blue dye prior to a total hysterectomy demonstrated good blue dye spread to the parauterine lymphatics, the area of uterine vessel drainage, a major lymphatic draining route from the uterus . The addition of the fundal injection did not appear to improve on the detection rate, especially when the cervical injection was satisfactory.
This technique is more complicated and demanding than the direct cervical or uterine corpus injection as it requires hysteroscopic evaluation. Multiple studies have used the technique of hysteroscopic injection to identify SLNs (Table 4) [13,23–26]. Niikura et al. , in their series of 28 patients, injected radioactive fluid around the tumor under direct visualization by hysteroscopy the day prior to surgery. The injections were performed at 4 sites under the endometrium around the tumor, and blue dye was used only to ensure the absence of leakage and not to identify SLNs. For patients with multiple or diffuse tumors in the uterine cavity, 5 injection sites covering the entire uterine cavity were used. Lymphoscintigraphy was performed 10 minutes later and again the next day prior to surgery. SLNs were identified intraoperatively by a gamma probe and removed. Pelvic and paraaortic lymphadenectomies were then performed along with a total abdominal hysterectomy and bilateral salpingo-oophorectomy. The detection rate was 82%, with 100% sensitivity and 100% specificity. The location of SLNs identified in this report appears similar to the location of metastatic nodes identified in other studies examining spread patterns in endometrial cancer, with an incidence of metastatic lymph nodes of approximately 50% in the pelvis, 30% in both the pelvis and the paraaortic area, and 20% in the paraaortic area alone . Maccauro et al.  and Raspagliesi et al.  evaluated both a radioactive tracer and blue dye injected hysteroscopically for the identification of SLNs. Using a slight modification of the technique described by Niikura et al. , they obtained a lymphoscintigraphy 15 minutes after the hysteroscopic injection and then every 5 minutes for up to 1 hour or until 2 or 3 SLNs were identified. Surgery was performed within 3–4 hours of the injection using an intraoperative gamma probe and direct identification of blue-stained nodes. After the removal of the SLNs, a pelvic lymphadenectomy was performed in all cases, and a paraaortic lymphadenectomy was performed only in cases of serous or clear cell histology. The authors reported that scintigraphy had a 100% success rate in detecting SLNs, while the success rate for blue dye was only in the 30% range in both studies. In addition, they reported that all metastatic nodes were part of the SLNs. In their series of 10 patients, on the other hand, Fersis et al. reported only a 50% detection rate . Gien et al.  used hysteroscopic injection of blue dye alone. In their first 3 patients, no SLNs were detected. In their following 4 patients, they used both subserosal and hysteroscopic injections and were only able to achieve a 50% detection rate. The reason for this lower detection rate is not clear but may be due to the small number of patients included in these last 2 studies [13,24]. One of the theoretical concerns when performing hysteroscopic injection in patients with endometrial cancer is the risk of disseminating malignant cells through the tubes. Hysteroscopic visualization of the endometrial cavity can be achieved with a pressure of 40 mmHg, which is lower than the 70 mmHg pressure needed for tubal spillage to occur . Maccauro et al.  and Raspagliesi et al., in their respective series of 26 and 14 patients, each reported only 1 patient with positive peritoneal cytology. Gien et al.  took peritoneal washings after the hysteroscopic procedure and found no cases of positive peritoneal cytology.
In summary, SLN mapping using a cervical injection with combined Tc and blue dye was feasible and accurate in patients with grade 1 endometrial cancer and may be a reasonable surgical option for this select group of patients. It is likely that this group of patients with endometrial cancer may be best suited for this procedure as this is the group that is more likely to have disease confined to the uterus at the time of surgery and less likely to have disruption of lymphatics with bulky disease that interfere with a successful, accurate identification of SLNs.
At present, regional lymphadenectomy remains the gold standard in many gynecologic oncology practices, particularly for the approximately 15% of cases with failed SLN mapping. However, there is still controversy regarding this practice, particularly in view of recent results from clinical trials . Our encouraging pilot results will hopefully stimulate further evaluation of SLN in grade 1 endometrial cancer; in addition, further evaluation of factors that may contribute to a failed mapping is necessary in a larger cohort of patients.
Obtaining satisfactory results requires a dedicated surgical, radiology, and pathology team to improve SLN detection rates. Further studies evaluating the role of IHC-only positive nodes would be warranted, should larger numbers of cases be identified, to determine their clinical significance in uterine cancers.
We would like to acknowledge and thank Lee K. Tan, M.D. for her contribution on the methodology of sentinel lymph node pathologic assessment and evaluation. We also would like to thank the gynecology service fellows and Physician assistants for their help in performing the radiolabeled injections.
CONFLICT OF INTEREST STATEMENT
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.