Combined preoperative prognostic nutritional index and D-dimer score predicts outcome in colorectal cancer

Colorectal cancer (CRC) is the third most lethal cancer in the world with a heavy health burden in China, where an estimated 592,232 new cases were diagnosed and up to 309,114 deaths occurred in 2022 [1]. Globally, the incidence of CRC has more than doubled with a corresponding remarkable increase in deaths noted between 1990 and 2019 [2]. Although an increasing number of patients are diagnosed and cured at a very early stage due to sigmoidoscopy screening [3], the 5-year survival rate for those with stage II and III disease remains unsatisfactory [4].

It is well established that the outcome of cancer patients is determined by many factors, including nutrition status [5] and anticancer immunity [6], in addition to the cancer cells themselves; accordingly, it is plausible that prognostic markers could be more reasonable if these factors are comprehensively considered. The prognostic nutritional index (PNI), which is a marker calculated based on serum albumin and lymphocytes, is a robust prognostic indicator in many malignancies, including lung cancer [7], nasopharyngeal cancer [8], liver cancer [9], breast cancer [10], gastric cancer [11] and CRC [12, 13]. Interestingly, previous studies have indicated that the prognostic efficacy of PNI was superior to that of other inflammation-based prognostic indicators [14, 15]. In CRC, the PNI was also noted to be a better marker than other inflammatory indicators and was the only independent risk factor for survival in stage IIA or III cases [16, 17]. Nonetheless, it was notable that PNI alone was limited by its prognostic efficacy. Specifically, the area under the curve (AUC) ranged from 0.56 to 0.67 in non-metastatic cases with a relatively low sensitivity (58.6%) and specificity (59.6–78.3%) [17, 18] and was only 0.62 in metastatic cases [19]. In recent years, some studies have reported a combination of PNI with other markers and found moderately improved AUCs in predicting the outcome. For example, the PNI was combined with the albumin-to-globulin ratio as a prognostic indicator of esophagogastric junction cancer [20] and with hemoglobin to predict the prognosis of esophageal squamous cancer [21]. In recent years, the pivotal role of circulating tumor cells (CTCs) in patients who relapse after surgery has become increasingly popular, and these cells are considered to be a strong prognostic indicator in stage I–III cases of CRC [22, 23]. Interestingly, a study reported that a combination of the controlling nutritional status score (an index calculated based on serum albumin, lymphocyte counts and total cholesterol that is similar to PNI) and CTCs potentially exhibits superior prognostic efficacy and is able to distinguish the survival of subgroups [24]. However, of note, the detection of CTCs was limited by specific technology and was not routinely conducted in practice; other alternative indicators are still needed.

D-dimer (DD), which is an end product in the biological process of fibrin degradation, is a sensitive indicator of coagulation and fibrinolysis. Interestingly, DD also has important value in many cancers [25‐27]. Similar to CRC, DD has long been established as a useful tumor marker, and its diagnostic value in preoperative staging is comparable to that of carcinoembryonic antigen (CEA) [28]. The prognostic value of DD in CRC has also been well studied [29‐31] and was demonstrated to be even better than CEA in metastatic scenarios [32]. More importantly, some previous studies indicated that the level of DD was closely correlated with CTCs in some cancers [33, 34]. Based on these facts, we hypothesize that a combination of PNI and DD could have good prognostic efficacy in CRC; however, related studies are limited.

Here, we aimed to explore the prognostic role of a combination of PNI and DD score (PDS) in CRC patients after curative surgery.

In this study, we found that the PDS was a useful prognostic marker for CRC patients after curative surgery. PDS 0 patients had the worst outcome, and OS could be well distinguished among PDS subgroups. To the best of our knowledge, this is the first report about the prognostic value of PDS in cancer.

The ultimate survival of cancer patients is determined by many factors in addition to the presence of cancer cells. PNI is a marker that combines nutritional status and anticancer immunity in patients, and it is plausible that its prognostic efficacy would be superior to that of other markers that only consider a single factor. In fact, some previous studies have validated this hypothesis. For example, Imai et al. assessed 717 consecutive hepatocellular cancer patients after curative resection and found that PNI was superior to controlling the nutritional status score, NLR, PLR and Glasgow prognostic score in both DFS and OS; PNI was the only independent risk factor in multivariate analyses [14]. Consistent with this finding, Komura et al. enrolled 308 epithelial ovarian cancer patients and found that pretreatment PNI was better than single platelet count in predicting DFS [15]. As noted for CRC, numerous studies have investigated the role of PNI and other systematic inflammatory prognostic indicators. For example, Sato et al. collected 72 stage II-III obstructive patients and found that PNI was the only independent risk factor for DFS and OS compared to NLR, LMR and PLR [38]; accordingly, Maruyama et al. included 197 stage IIA patients and found that PNI was the only independent risk factor in contrast to NLR and PLR [16]. Nonetheless, PNI alone is still limited by its relatively low prognostic efficacy, as mentioned previously [17‐19], and could be further improved when some factors that reflect cancer cell features are taken into consideration based on our hypothesis. With the exception of direct quantification of the CTCs, the tumor markers were a good label of malignant cells. CEA is a classical tumor marker that is mainly released by colorectal cancer cells [39, 40] and is a good indicator of the aggressiveness of these cells [41, 42]. Previously, some investigators have tried to combine PNI with CEA to improve the prognostic efficacy. For example, Uejima et al. included 135 stage II patients and found that a combination of PNI and CEA exhibited good prognostic efficacy, but the 5-year DFS rates in the CEA^low/PNI^high, CEA^high/PNI^high and CEA^low/PNI^low groups were 100%, 100% and 97.4%, respectively, making it difficult to distinguish the survival differences [43]. Similarly, Xu et al. enrolled 513 stage II–III patients and reported that PNI and CEA represent a superior combination, but the 1- and 3-year OS rates in the aforementioned patient subgroups were undistinguishable [44]. Taking into account these results, it was suggested that an additional indicator that was more reliable than CEA was needed. DD has been consistently reported as a tumor marker in CRC [28] and is also a robust prognostic predictor in CRC [29, 30]. Interestingly, the AUC of DD was higher than that of CEA (0.85 vs. 0.72) with a significantly better sensitivity (88.0% vs. 65.2%) in CRC patients after curative resection [45]. Although no related reports have described the use of DD in combination with other prognostic makers in CRC, one study indicated that a combination of DD with NLR could be a useful prognostic indicator in non-small cell lung cancer (NSCLC), and the 5-year OS rates were distinguishable among the subgroups (23.5% vs. 34.2% vs. 50.0%) [46]. In our study, the 3-year OS rates in the PDS 0, 1 and 2 subgroups were 95.35%, 75.51% and 59.62%, respectively (P < 0.001), and could be effectively separated. However, when patients were divided into 4 subgroups (PNI^high/DD^low, PNI^high/DD^high, PNI^low/DD^low and PNI^low/DD^high) as noted in previous studies (only those stage II–III) [43, 44], the 3-year OS rates were 95.35%, 80.95%, 71.43%, and 58.82%, respectively (P = 0.001, data not shown), and it was also difficult to separate the survival differences among the subgroups except for the comparisons between the PNI^high/DD^low and PNI^low/DD^high as well as the PNI^high/DD^high and PNI^low/DD^high subgroups. These results were partly consistent with previous studies [43, 44].

Interestingly, it was proposed that cancer is actually a stem cell disease [47]. Colorectal cancer stem cells (CCSCs) identified by specific surface markers, such as CD44 and CD133, have been extensively studied [48]. In a previous study, CCSCs were thought to be the ultimate reason for cancer initiation, dissemination and recurrence, and eradiation of these cells represents a key approach to cure the disease [48, 49]. Of note, cancer dissemination could occur at the very beginning in CRC [50], and these cells in the circulating system (CTCs) exhibit some features of CCSCs [51]. In addition, cancer-promoted inflammation also plays a key role in determining CRC development [52], and inflammatory cytokines, such as interleukin-6 (IL-6) and IL-1, are significantly elevated in CRC patients [53‐55], which could have an important role in supporting CCSCs [56, 57]. Interestingly, albumin synthesis could be significantly decreased under inflammatory conditions, and IL-1 was postulated to be an important mediator [58]. Additionally, IL-6 contributes to the decreased expression of albumin genes through the activation of tyrosine kinase [59]. Lymphocytes are the major players in adaptive anticancer immunity [60] and specifically recognize and eradicate CCSCs [61]. Although not reported in CRC, it was found that the proportion of circulating lymphocytes could be decreased under inflammatory conditions [62, 63] in addition to the abnormal immune-suppression function induced by these cytokines [64]. In recent years, accumulating evidence has indicated that CTCs are the key source of relapse in CRC [65, 66]. Interestingly, DD is correlated with CTCs and is an essential accompaniment of these cells [67]. More importantly, the level of DD correlated with CTCs in breast cancer and non-small cell lung cancer [33, 34]. Based on these facts, it would be plausible that patients with PDS 0 have relatively low PNI and high DD levels that are equal to strong cancer-promoting inflammation and sustained high counts of CTCs, which would correlate with poor outcome. However, studies that directly and concurrently analyze the levels of inflammatory cytokines and PNI and DD remain lacking.

Clinically, adjuvant chemotherapy was conventionally recommended for high–risk stage II and stage III CRC patients [68, 69], which could greatly improve the DFS and OS [68‐71]. Previously, pT4, lymph-vascular or perineural invasion, perforation or obstruction presentation, poorly differentiated histology, or lymph node harvest less than 12 were well acknowledged as risk factors in CRC [72]. In our study, PDS was found to be useful in prognostic prediction and PDS 0 patients displayed significant poor DFS and OS compared to other subgroups, we speculate maybe it could be considered as an additional risk factor for decision-making approach for the patients in practice; however, randomized controlled trails are necessary to confirm our speculation. In addition, taking into consideration that PDS was an easily accessible indicator, it may also valuable in monitoring the treatment response of metastatic disease to systemic therapies like CEA [73]; however, more studies are needed to validate its role in such a scenario.

There are some limitations to our study. First, it was a retrospective study with a relatively small sample size, and potential biases cannot be excluded. Second, information on adjuvant chemotherapy was insufficient, and it was notable that both the albumin levels and absolute lymphocyte counts could be altered by these therapies [74, 75]; thus, long-term PDS measurements and assessment of its prognostic value in these patients should be performed in the future.

Overall, our study suggested that PDS was a useful prognostic indicator in CRC patients after curative surgery and that PDS 0 patients exhibit inferior survival. Additional studies are needed to validate these findings in the future.