Long-Term Outcomes of Surgical Resection of Pathologically Confirmed Isolated Para-Aortic Lymph Node Metastases in Colorectal Cancer: A Systematic Review

Simple Summary Para-aortic lymph node (PALN) metastases represent patterns of initial recurrence in only 2–6% CRC patients. Lack of consensus has impaired an unambiguous statement for PALN recurrence resection. Our systematic review identified 59.4–68% 3-year OS rate and 53.4–87.5% 5-year OS rate, with a 25–84 months median OS, 26.3–61% 3-year DFS rate and 0–60.5% 5-year DFS rate, with a 14–24 month median DFS, in patient undergoing isolated PALNM resection. Overall, 62.1% re-recurrence rate ranged from 43.8% to 100%. Although PALNMs resection in CRC patients may be considered a feasible and beneficial option, no conclusions or recommendations can be provided, taking into account the current evidence. Further randomized, possibly multicenter trials are strongly recommended and mandatory in order to confirm our results and clearly identify patient selection criteria. Abstract Background: Para-aortic lymph node (PALN) metastases represent patterns of initial recurrence in only 2–6% CRC patients, after an estimated 23–28 month time interval. An increasing trend towards curative surgery has been witnessed in patients presenting with controlled PALN recurrence. Nevertheless, lack of consensus has impaired an unambiguous statement for PALN recurrence resection. Methods: We performed a systematic literature review following the Preferred Reporting Items for Systematic Reviews and Meta-Analyzes (PRISMA) guidelines, which led us to gain deeper insight into the prognostic factors and long-term outcomes after resection for synchronous or metachronous pathologically confirmed CRC isolated para-aortic lymph node metastases (PALNM). Pubmed/MEDLINE, Embase, Scopus, Cochrane Library and Web of Science databases were used to search all related literature. Results: The nine articles included covered a study period of 30 years (1988–2018), with a total of 161 patients. At presentation, most primary CRCs were located in the colon (74%) and 95.6%, 87.1% and 76.9% patients had T3–T4, N1–N2 and well/moderately differentiated CRC, respectively. We identified a 59.4–68% 3-year OS rate and 53.4–87.5% 5-year OS rate, with a 25–84 months median OS, 26.3–61% 3-year DFS rate and 0–60.5% 5-year DFS rate, with a 14–24 month median DFS. Overall, 62.1% re-recurrence rate ranged from 43.8% to 100%. Conclusions: Although PALNMs resection in CRC patients may be considered a feasible and beneficial option, no conclusions or recommendations can be made taking into account the current evidence. Therefore, further randomized, possibly multicenter trials are strongly recommended and mandatory if we want to have our results confirmed and patient selection criteria clearly identified.


Introduction
Colorectal cancer (CRC) is the third most commonly occurring cancer, which has become the second leading cause of cancer-related deaths worldwide [1,2]. In CRC patients, metastatic dissemination represents the most frequent cause of death, as approximately 25% patients show CRC metastases at diagnosis [1].
As approximately 50% patients develop hepatic metastases in the course of the disease, the liver represents the most common site of CRC metastases [1]. About 20% of patients affected by hepatic metastases show a potentially resectable disease [1]. Following resective surgery of isolated hepatic metastases, low survival rates were reported (25-58% at 5 years, 17-28% at 10 years) [1], and the lung ranks second (10-15% patients with CRC at diagnosis) [1]. After the resection of isolated pulmonary metastases, survival rates range between 32% and 68% at 5 years and between 11% and 34% at 10 years [1].
Although CRC hepatic and pulmonary metastases require standardized management, how to treat lymph node recurrence (either isolated or combined with other CRC metastatic lesions) remains an open question [1][2][3]. Some authors have made a distinction between retroperitoneal local recurrence and RLN recurrence [3]. Moreover, although complex protocols have been suggested, none of them have been fully validated and adopted [3]. An increasing trend towards curative surgery has been witnessed in patients presenting with controlled RLN recurrence [3]. Nevertheless, lack of consensus impaired an unambiguous statement for RLN recurrence resection [3].
A thorough systematic review of literature led us to achieve deeper knowledge of prognostic factors and long-term outcomes after resection for synchronous or metachronous pathologically confirmed CRC isolated para-aortic lymph node metastases (PALNM).

Search Strategy
We performed a systematic literature review in compliance with Preferred Reporting Items for Systematic Reviews and Meta-Analyzes (PRISMA) guidelines. According to the gold standard for literature search related to surgical reviews [4], the following databases were searched: PubMed/MEDLINE, Embase, Scopus, Cochrane Library (Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials-CENTRAL) and Web of Science (Science and Social Science Citation Index). Combination of non-MeSH/MeSH terms was as follows: -PubMed/MEDLINE (Para-aortic OR Paraaortic) AND (((("Colorectal Neoplasms" [Mesh])) OR ((Colorectal OR Colon OR Rectal) AND (cancer OR tumor OR tumour OR neoplasm))) AND (("Lymph Nodes" [Mesh]) OR "Lymph Node Excision" [Mesh] OR Lymph node* OR Lymphadenectom* OR Lymph nodes dissection*)) Filters: English, Italian.

Terminology
To reach a better understanding of results, we adopted the following descriptions: RLN recurrence: histologically confirmed retroperitoneal lymph node metastasis/es, without local CRC recurrence, in an area which is laterally limited by the ureters, superiorly limited by celiac area and inferiorly limited by iliac vessels [5].
PALNM: histologically confirmed recurrence in lymph nodes surrounding abdominal aorta and inferior vena cava, with lateral extension to edge of the psoas major muscles, superior extension to the crura of the diaphragm and caudal extension to mid common iliac vessels [6].
RLN recurrence and PALNM were adopted by Authors to define the same pathological condition.
PALNM type A: A1-Lymph node metastases located in the area of the aortic hiatus, about 4-5 cm in width surrounded by the medial crus of diaphragm (these nodes are located within median arcuate ligament of diaphragm); A2-Lymph nodes metastases located in the area from the uppermost part of the origin of the celiac trunk to the lower margin of the left renal vein [6].
PALNM type B: B1-Lymph node metastases located in the area from the lower margin of the left renal vein to the uppermost part of the origin of the inferior mesenteric artery; B2-Lymph nodes metastases located in the area from the upper margin of the origin of the inferior mesenteric artery to the aortic bifurcation [6].
Isolated PALNM: Lymph node metastases when there are no CRC metastatic sites other than PALNs [7].
Combined CRC metastases: Lymph node metastases when one or more sites have CRC metastases other than PALNs [8].
PALN dissection (PALND): separation and removal of lymphatic tissue from the para-aortic vessels.

Inclusion Criteria
Our search was limited to articles published from January 2000 to February 2021. The selection included just scientific papers in English or Italian language.
Our systematic review covered population studies (case series, case-control studies, cohort studies, controlled clinical trials and randomized clinical trials) which included 5 or more adult patients (over 18 years of age), who underwent surgical resection for pathologically confirmed CRC isolated PALNM. We ruled out abstracts, posters, case reports, previously published systematic reviews and/or meta-analyses and population studies analysing less than five patients.
Our review ruled out articles that examined mixed clinical populations, isolated PALNM plus combined CRC metastases, or exclusively combined CRC metastases populations, as well as mixed pathologically populations (pathologically positive plus pathologically negative PALNM) after PALND.
In addition, articles that reported or led to data on survival (overall survival-OS-rate and/or median survival) and disease-free survival (disease-free survival-DFS-rate and/or median disease-free survival) were included, and we ruled out articles that did not lead to data related to survival and disease-free survival. Moreover, reference lists of included studies were manually scanned to further identify any potentially relevant articles.

Data Extraction
The papers were selected and identified by two independent reviewers (M.Zi. and M.C.B.) based on title, abstracts, keywords and full-texts. The following information was collected from the papers used in this study: demographic data [author's surname and year of publication, study period, study type, study Center, population size, gender and age, primary CRC location, CEA and CA 19.9 levels at PALNM resection, chemotherapy (CT) and radiotherapy]; histopathological data [primary CRC T and N stage, primary CRC histology]; PALNM data [timing of presentation, anatomical location, PALN harvested, PALNM retrieved]; outcomes data [mortality rates, median overall survival, disease-free survival rates, median disease-free survival, follow-up duration, recurrence]. All collected results were eventually reviewed by a third independent reviewer (M.P.F.D.).

Quality Assessment
The Newcastle-Ottawa quality assessment scale (NOS) was used to assess the quality of each study [9]. The thresholds for converting the Newcastle-Ottawa scales to AHRQ standards are as follows (good, fair, and poor): (i) good quality: 3 or 4 stars in selection domain AND 1 or 2 stars in comparability domain AND 2 or 3 stars in outcome/exposure domain, (ii) fair quality: 2 stars in selection domain AND 1 or 2 stars in comparability domain AND 2 or 3 stars in outcome/exposure domain; (iii) poor quality: 0 or 1 star in selection domain OR 0 stars in comparability domain OR 0 or 1 stars in outcome/exposure domain.

Search Results and Study Characteristics
A final literature search conducted in February 2021 helped to identify 729 studies of potential interest [PubMed/MEDLINE: 217 records; Embase: 120 records; Scopus: 170 records; Cochrane Library: 12 records; Web of Science: 210 records] ( Figure 1). One additional record was found using another source (References list).

Discussion
Lymph nodes represent one of the most commonly affected sites of metastasis [18]. Tumours spread to lymph nodes via lymphatic vessels and this usually happens in order of proximity to the primary site [18]. Lymph node (N category) staging includes information on whether cancer has spread to regional lymph nodes and on the number of involved lymph nodes [18]. Regional lymph nodes, which are located along the course of major vessels that supply the large intestine, are designated based on the anatomical subsite of large bowel [18]. Lymph nodes, which are located outside the regional drainage area of a primary tumour, should be described as distant metastases (M category) [18].
With an incidence rate of less than 2%, PALN involvement in CRC is uncommon and American Joint Committee on Cancer (AJCC) considers it to be a disseminated stage IV disease, although Japanese Society for Cancer of the Colon and Rectum (JSCCR) consider regional PALNM to be a stage III disease [19,20].
Synchronous isolated PALNM was found to be 3.4% for right colon cancer, 3.9% for left colon cancer and 3.1% for rectal cancer [7]. Recording a 1.0-1.3% range, metachronous isolated PALNM is a rare condition, as PALNM is frequently followed by other distant site metastases [7]. In CRC, isolated PALNM has not been classified and it has been previously classified as retroperitoneal recurrence, a kind of locoregional recurrence [7]. Retroperitoneal recurrence does not only include PALNM, but also the growth of a tumor deposit or residual tumour from surgery [7]. Thus, isolated PALNM must be carefully investigated in terms of previous reports, in order to identify its management in the field of CRC [7] As is the case for PALNM resection from other gastrointestinal tract cancers (e.g., gastric or pancreatic cancer), some authors have stressed that CRC PLANM resection can lead to better oncological outcomes in selected patients [21][22][23]. Indeed, some studies considered RLN metastases to be an extension of mesenteric lymph node metastases, or second-echelon regional lymph node metastases, with the potential for curative resection [5,[10][11][12][13][14][15][16][17]. From this perspective, the involvement of major vessels in the retroperitoneum represents the only barrier for curative resection [16].
Three previously published systematic reviews had investigated the same topic we chose [7,8,24]. In 2011, Ho et al. studied a 110 patient population (isolated plus combined; surgery alone), identifying a 34-44 month median OS and a 17-21 month median DFS [24]. In 2016, Wong et al. investigated a patient population of 370 (isolated alone; surgical plus no surgical group); out of whom, 145 patients had synchronous metastases and 225 ones had metachronous metastases. Furthermore, 8 5-year OS and DFS seemed to be relatively similar; in the synchronous PALNM group, 5-year OS ranged from 22.7% to 65.7%, while DFS ranged from 17.6% to 40.2%; in the metachronous PALNM group, 5-year DFS ranged from 15% to 60%, while DFS ranged from 10% to 25.6% [8]. In 2018, Sasaki et al. analyzed an overall population of 227 patients (isolated plus combined; surgical plus no surgical group) [7]. The 3-year OS ranged from 60% to 100%, with a median 34-80 month OS for patients who underwent PALND [7]. However, findings from the aforementioned studies differ from our analysis, as they also took into account case reports or populations with less than five patients but, above all, also included patients who had undergone PALND based on radiological findings, although reported to be pathologically negative later.
The potential impact of a time lapse to presentation of metastases in PALNM patients represents a highly interesting topic. According to the literature, synchronous metastases in primary CRC tend to record a worse prognosis if compared to metachronous presentation [1]. Ruling out the study by Choi et al., which included a mixed synchronous/metachronous population, we recorded a 34-84 month median OS in a synchronous PALNM population and a 25-37 month median OS in metachronous PALNM [5,[10][11][12][13][14][15][16][17]. On the other hand, DFS in both populations were recorded at similar rates (17-20.3 months in synchronous PALNM population versus 22-24 months in metachronous PALNM one) [5,[10][11][12][13][14][15][16][17]. However, with regard to patient population heterogeneity, in addition to patient selection bias and different cancer biology (e.g., right colic cancer versus left colic cancer versus rectal cancer), the above mentioned results should be considered with caution.
In the selection of patients suitable for PALND, Albandar et al. emphasized the role of perioperative chemotherapy or chemoradiotherapy (CRT), as tumor regression or nonprogression of other metastases after CT or CRT may mean a favourable biology, thereby identifying patients who are more likely to benefit from surgery, i.e., very few patients who undergo preoperative treatment and almost all those who have postoperative treatment [20]. However, none of the abovementioned studies provided an indication of pre-and postoperative CT/CRT or their consequences on subsequent surgical management [5,[10][11][12][13][14][15][16][17]. Furthermore, only five of the nine authors reported the chemotherapy regimens they had used [10][11][12][13][14]. In particular, we identified eight different chemotherapy regimens in the absence of plain reasons for their indication (see Supplementary material- Table S2). Among the five abovementioned authors, only Ogura et al. analyzed impact of chosen chemotherapy regimen on survival: they identified an absence of statistical significance between no oxaliplatin/CPT-11 group and oxaliplatin/CPT-11 one (55.6% vs. 77.8% 5-year cancer-specific survival) [13]. Of the 9 studies, just Kim et al. reported patients (4) also undergoing adjuvant RT [16]. However, the authors did not describe both the RT schemes adopted and the possible influence of RT on survival [16].
Non-surgical management (CT and/or radiation therapy-RT) has been used as rescue therapy for PALNM patients.
Unfortunately, it was not possible to obtain details about the potential relevance of the resection margin on long-term outcomes. However, we believe that the resection margin status could play an important role.
Until now, we have focused our attention on a detailed presentation of the results related to our review on oncological outcomes and prognostic factors after CRC PALNMs resection. Although it does not represent the main purpose of our study, we think it is useful to complete our manuscript with the presentation of two interesting surgical endpoints, namely, indications adopted for PALND and description of postoperative complications.
Finally, we tried to identify both the number and type of complications following PALND (see Supplementary material- Table S4). However, only five out of the nine included studies reported such information and recorded a 19-47% rate of complications, thus perfectly in line with previous reports in the scientific literature [5,10,11,13,17].

Limitations
Our systematic review has some limitations, including: (i) the literature search did not include non-English-written scientific papers, except for studies in the Italian language; (ii) reported events were mainly part of small retrospective series; (ii) populations under analysis showed heterogeneity; (iii) the study time frame witnessed variation of diagnostic methods, surgical techniques and skills, perioperative CT/CRT protocols; (iv) many relevant data were not thoroughly described by the Authors, as reported in Tables 1-5; (v) many data were reported through median or mean results. All these reasons made direct comparison of findings arduous.

Conclusions
Currently, an agreement on the best treatment for patients with isolated PALNM from CRC has not been reached. Curative CRT alone may represent an option for poor surgical patients or those patients who show a poor performance. Preoperative CT or CRT followed by surgery should be taken into account in superselected patients who are likely to benefit from PALNM resection. However, although PALNMs resection in CRC patients may be considered as a feasible and beneficial option, no conclusions or recommendations can be made taking into account the current evidence.
Moreover, recurrence rates after PALNM resection seem to be high. Therefore, further randomized, possibly multicenter trials are strongly recommended and mandatory if we want to confirm our results, and patient selection criteria need to be clearly identified.
Supplementary Materials: The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/cancers14030661/s1, Table S1: Newcastle-Ottawa Quality Assessment Form for Cohort Studies; Table S2: Chemotherapy regimens of included studies; Table S3: Indications adopted for PALND; Table S4 Institutional Review Board Statement: Local ethics committee (Comitato Etico dell'Area Vasta Emilia Nord, Italy) ruled that no formal ethics approval was required in this study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.