At present, aggressive hepatic or pulmonary resection is the standard strategy for managing liver or lung metastasis with significant better prognosis when R0 resection can be achieved. As to PALN metastasis, however little is known. The reason for this may be due to the relatively low incidence of PALN metastasis and its difficulty of diagnosis. The reported incidences of PALN metastasis in left-sided CRC from three large case series are 0.93%-3.41% [6,12,17]. Accurate estimate of the rate of PALN metastasis is impossible, since confirmation of PALN metastasis rely on pathology. Because of the complexity and complication of PALN dissection [15], PALN dissection was not routinely performed and evaluation of its indication depends on pre-operative radiographic examination. In addition, a considerable number of patients with advanced tumors diagnosed by radiographic methods cannot be treated with surgical operation, and it impossible to confirm the diagnosis of PALN metastasis in these patients. On the other hand, only 56–60% patients could achieve R0 resection for PALN metastasis, and even in these patients, only a subset of patients could benefit from PALN dissection [6,17-19]. Therefore, it is important to improve the diagnosis accuracy and screen out patients with potential benefit though radiographic and other preoperative clinical data. In addition, for the patients who have not received PALN dissection, it is necessary to screen out the patients with high-risk of PALN metastasis to receive targeted postoperative chemoradiotherapy for better prognosis.
However, there is currently no accepted standard for the CT diagnosis of PALN metastasis, and the diagnostic criteria varied among different studies, resulting in a large difference in the incidence of radiographic PALN metastasis, ranging from 3.0 to 14.3%, while the positive predictive value CT is only 19.3-32.3% [6,11,12,17]. On this account, our study adopted a relatively strict diagnostic standard, that is, the lymph node with the short-axis diameter of ≥ 10mm was regarded as PALN metastasis. The incidence of radiographic PALN metastasis in our study was 3.56%, which is lower than the above-mentioned data, but comparable to their incidence of pathological PALN metastasis [6,17]. Even so, the accuracy of a single CT diagnosis is still not satisfactory, and it is necessary to combine the clinical data to improve the diagnostic accuracy. Our results showed that the pre-treatment clinical data including serum CEA, CA19-9 levels, radiographic liver and IMLN metastases, among them, radiographic IMLN metastases was an independent risk factor for radiographic PALN metastasis.
Serum tumor markers play an important role in prognosis prediction for CRC patients due to their benefit-cost ratio and convenience of measurement. CEA is recommended by main guidelines to be routinely measured in CRC [20,21]. Elevated preoperative serum CEA levels was associated with PALN metastasis and poor survival [6,16]. Although not recommended in CRC guidelines, CA19-9 is a commonly used tumor marker in pancreatic cancer, as well as several other kinds of gastrointestinal cancers [22,23]. Preoperative CA19-9 is an independent prognostic factor in CRC patients, and the patients with elevated CA19-9 have shorter overall survival, disease-free survival and recurrence-free survival [24,25]. In our study, the incidence of PALN metastasis in the patients with CEA level of ≥ 5ng/ml or CA19-9 level of ≥ 37U/ml was significantly higher than that in the patients with normal CEA or CA19-9 level, indicating that elevated CEA or CA19-9 level was a risk factor for radiographic PALN metastasis.
The liver is the most common site of distant metastases in CRC, and about 14.5% -18.9% of CRC patients have synchronous liver metastases at initial diagnosis [26,27]. The correlation between lymph node and liver metastasis in CRC is still controversial [28,29]. But our study showed that the incidence of PALN metastasis in the liver metastases group was significantly higher than that in the group without liver metastases. However, since liver metastases are mostly from hematogenous spread, they may be not specific to PALN metastasis, but rather, they are both the indicators of advanced CRC.
Lymph node metastasis is one of the most important and reliable prognostic factors in CRC and is staged with the number and distribution of involved lymph nodes. According to the 8th AJCC staging manual, IMLN, the last barrier against PALN metastasis, is categorized as regional lymph node in left-sided colon and rectal cancer [7]. Previous studies suggested that 32.0%-59.5% of CRC cases with IMLN metastasis had PALN metastasis, and IMLN metastasis is an independent risk factor for PALN metastasis [6,12]. In our study, the incidence rate of PALN metastasis in the patients with IMLN metastasis was 31.1%, and unlike other risk factors, IMLN metastasis was an independent risk factor for PALN metastasis in multivariate binary logistic regression analysis.
With regard to pathological characteristics associated with the presence of PALN metastasis, as with the previous researches [6,12], the relatively longer tumor length, lower differentiation type, more advanced TNM stage and lymphovascular or perineural involvement were risk factors for radiographic PALN metastasis in our study. But in multivariate binary logistic regression analysis, it was failed to demonstrate that any of them was an independent risk factor for radiographic PALN metastasis. Thus, including liver metastasis, elevated CEA and CA19-9 level, each of these non-independent risk factors shown in PALN metastasis patients might be not specific to PALN metastasis, but rather, they all were the common characteristics of almost all types of advanced CRC and were the risk factors for poor prognosis [6,12,17].
Currently, there is no consensus over the management for CRC with PALN metastasis. Some previous researches showed the potential beneficial of multidisciplinary treatment in carefully selected patients, including PALN dissection and preoperational chemoradiotherapy [6,8-12,30]. Synthesizing the results of our study, we recommend PALN dissection for the patients with visible PALNs measuring ≥10 mm in the short axis or IMLN metastasis confirmed by intraoperative pathological biopsy. For the patients with higher risk, including synchronous liver metastasis, elevated CEA or CA19-9 levels, preoperative conversion chemoradiotherapy may be preferred. For the patients without radiographic PALN metastasis and have not received PALN dissection, if pathological results of primary tumor show associated risk factors and chemotherapy is originally indicated after surgical operation, a relatively high-intensity chemotherapy regimen is recommended. In addition, on the basis of the postoperative follow-up protocol recommended by the guidelines, the frequency of CT examination should be increased to detect PALN metastasis as early as possible.
There were some limitations in our study. First, this was a retrospective study at a single center with a small sample size and selection bias. Second, because PALN resection was not routinely carried out, it was difficult to estimate the incidence of PALN metastasis in real world, so we take radiographic PALN metastasis as instead, thus false positive or negative case was inevitable. To evaluate the efficacy of CT diagnosis accurately, a randomized clinical trial will be necessary.