Colorectal cancer (CRC) is the third most common cancer and has the second highest cancer-related mortality rate (1); postoperative adjuvant chemotherapy is the standard treatment in stage III CRC (2). Oxaliplatin-based chemotherapy is a key regimen for CRC, and oxaliplatin is included in the capecitabine plus oxaliplatin (CapeOX) regimen (3). However, oxaliplatin may induce hepatic sinusoidal obstruction syndrome (SOS) (4).

Approximately 30% of patients with CRC develop metachronous liver metastases, and hepatectomy is the only potentially curative treatment (5). Moreover, approximately half of patients treated with oxaliplatin develop SOS, which leads to significant postoperative adverse effects, especially after major liver resection (6-8). Early SOS assessment is necessary; however, an effective strategy remains to be determined.

Increasing splenic volume (SV) may predict the risk of SOS (9); however, measuring SV is complicated. SOS may result from drug-induced injury to liver sinusoidal endothelial cells (LSEC), and hyaluronic acid (HA), present in LSEC, was a marker for SOS in a rat experimental model (10). Therefore, we evaluated HA as a prognostic marker for early SOS assessment. We measured HA before and after CapeOX adjuvant chemotherapy in patients with stage III CRC after curative surgical resection, and investigated the relationships and changes in treatment progression between SV, platelets, and liver dysfunction, as indicators of SOS.

Patient selection. We identified 34 patients with stage III CRC in our hospital database who underwent radical resection between January 2017 and June 2020, and who received CapeOX adjuvant chemotherapy. We excluded patients with <6 oxaliplatin-based adjuvant chemotherapy cycles; lacking blood HA results or computed tomography (CT) imaging; metastatic disease during chemotherapy; additional chemotherapy agents or prior splenectomy; known cirrhosis; chronic viral hepatitis; or prior liver surgery. This study was approved by the Institutional Review Board of the Kanazawa Medical University Hospital.

Clinical data. We retrospectively collected the patients’ sex, age, body mass index, characteristics of the primary cancer and surgical procedure, postoperative course, number of chemotherapy cycles, pathological factors, SV, and laboratory data, namely hepatobiliary system enzymes, platelets, and HA concentration. SV and laboratory data were measured four times: i: before, ii: immediately after, iii: 6 months after, and iv: 1 year after adjuvant CapeOX therapy. Changes in SV were determined by comparing the value at each time point with the value before treatment. In this study, we defined splenomegaly as a 30% increase in SV when comparing ii to i. The correlations between HA change and SV change ratio, platelets, and liver enzymes were investigated. The patients were divided into two groups: Group A: ≥30% increase in SV at ii compared with i; Group B: the remaining patients. The outcomes for all patients and for each group were investigated. Thrombocytopenia was defined as <150,000 platelets/mm³. SV was calculated using multi-detector row CT (Hitachi Medical Corporation, Saitama, Japan) and by loading the CT images onto the calculator software, Ziostation 2^® (Zaiosoft, Tokyo, Japan). Additionally, we performed hepatectomy for recurrent liver metastasis in two patients in each group >6 months after CapeOX, and pathologically examined a normal site in the hepatectomy specimen and investigated the relationship between the pathological findings and preoperative HA level.

Statistical analysis. Values were expressed as means±standard deviations (SD). Statistical analyses were performed using the two-sided Student’s t-test and the Mann–Whitney U-test for continuous data or Fisher’s exact test. All statistical analyses were performed using SPSS 10.0 (SPSS, Chicago, IL, USA). Significance was defined as p<0.05.

Thirty-four patients were included; Group A (splenomegaly) comprised 17 patients, and Group B (no splenomegaly) comprised 17 patients. There were no differences in the groups’ clinical characteristics (Table I).

Analyzing all patients, adjuvant CapeOX resulted in a significant increase in HA at ii vs. i (p<0.001), and at iii (p=0.020). Platelet counts decreased significantly at ii (p<0.001), which persisted at iii (p=0.002) and iv (p=0.011). Aspartate aminotransferase (AST) (p<0.001) and alanine aminotransferase (ALT) (p=0.007) levels increased significantly at ii compared to i; AST levels remained significantly increased at iii (p=0.001) (Figure 1A). Changes in SV for all patients are presented in Figure 2. The mean SV at ii increased significantly to 137.3% compared to i (p=0.032). Resolution of splenomegaly was confirmed (mean SV: 121%) at iii, and SV returned to almost the initial level (mean: 110.1%) at iv (Table II). The percentage change in SV from i to ii was related to the percentage change in HA from i to ii (Figure 3).

Changes in laboratory data within in the groups are shown in Figure 1B and C. In Group A, HA levels increased significantly at ii (p<0.001) and iii (p=0.013) compared with i. Platelet counts decreased significantly at ii (p<0.001), which persisted at iii (p=0.009) and iv (p=0.018). AST (p<0.001), ALT (p=0.007), and γ-glutamyl transpeptidase (γGT) (p=0.029) levels increased significantly at ii compared to i. AST (p=0.002) and γGT (p=0.021) levels remained significantly changed at iii (Figure 1B). The mean SV at ii and iii increased significantly to 163.4% (p=0.009) and 136.5% (p=0.061), respectively, compared to i, and improved to 119.4% at iv (Table II). However, SV in three patients (21%) remained ≥30% even at iv (Figure 2). In contrast, in Group B, HA levels increased significantly at ii (p=0.014), then improved at iii and iv. Platelet counts decreased significantly at ii (p=0.008), and improved at iii and iv. AST levels increased significantly (p=0.004) at ii compared with i; however, values improved, and did not change significantly at iii (Figure 1C). Mean SV did not increase at any time after chemotherapy, compared to i (Table II). Thrombocytopenia (cut-off: <15 × 104/μl) during CapeOX treatment differed significantly in all 17 cases (100%) in Group A and in 4 cases (28.6%) in Group B (p<0.001).

Further examination of the HA results revealed that the cut-off value at time ii according to the receiver operating characteristic curve between Groups A and B was 138 ng/ml, with an area under the curve of 0.830 (Figure 4). HA in all patients in Group B was <138 ng/ml at ii, with a slightly low specificity of 64.7%, and sensitivity of 100%.

Of the 34 patients, 2 patients experienced liver recurrence >6 months after finishing chemotherapy. One patient in Group A underwent hepatectomy 4 months after CapeOX therapy, and another patient in Group B underwent hepatectomy 6 months after CapeOX. In the Group A patient, HA was 246 ng/ml at ii, and liver SOS was revealed pathologically (Figure 5A). However, in the Group B patient, HA was 94 ng/ml, with a normal liver (Figure 5B).

The novel findings in this study were that 50% of patients who received CapeOX as adjuvant chemotherapy for stage III CRC developed splenomegaly immediately after chemotherapy, and splenomegaly was associated with a significant increase in HA. Moreover, patients with significantly elevated HA had higher rates of remarkable thrombocytopenia and liver dysfunction during and even after CapeOX administration, indicative of SOS.

Although Rubbia-Brandt et al. (4) reported an association between oxaliplatin-based chemotherapy and SOS in patients with liver metastatic CRC, clinically useful prediction methods for this side-effect are poorly documented. Recent studies have addressed the efficacy of increasing SV to predict SOS in CRC patients (9,11). However, monitoring SV cannot be performed without CT, and CT imaging usually has monthly intervals. Recently, patients receiving CapeOX after CRC surgery (3) have developed SOS, which can be detrimental to the treatment of recurrent liver metastasis (6-8).

Several studies have reported that initial pathophysiological changes in SOS may result from drug-induced injury to LSEC (10,12,13), and HA is metabolized almost exclusively by LSEC (14). Serum HA is elevated after bone marrow transplantation in patients with moderate or severe SOS (15). Moreover, SOS in a monocrotaline-induced rat model of SOS showed that HA was significantly increased in severe SOS and significantly decreased in mild SOS (10,13), supporting HA as a predictive biomarker of SOS.

We found a strong correlation between patients with significantly elevated HA and those with a ≥30% increase in SV after CapeOX. In these patients, there was also marked and persistent thrombocytopenia, persistent liver dysfunction, and a persistent increase in HA compared to patients without increased HA during and after chemotherapy. There is a strong correlation between splenomegaly and thrombocytopenia or increased liver function caused by SOS in patients receiving oxaliplatin-based-chemotherapy (4,6,8,9); however, our results suggest that SOS evaluation is possible using HA instead of SV. Oxaliplatin-induced SOS may continue for more than a year after chemotherapy (9,11), consistent with our results.

The results of this study are important because they support the hypothesis that systemic indicators of drug-induced injury to LSEC may serve as simple biomarkers of SOS in CRC patients receiving oxaliplatin-based chemotherapy. However, in order to prove the specificity of measured HA levels for oxaliplatin-induced injury of liver sinusoidal endothelial cells, further studies are required that take into account the tumor as a source of systemic HA.

As a limitation, we evaluated the records of only 34 patients who received CapeOX chemotherapy for stage III CRC. Additionally, this was a retrospective and non-randomized study. Third, because liver biopsy was not performed, we chose increased HA as a possible biomarker of SOS. The correlation between liver dysfunction induced by chemotherapy and the development of SOS in pathological examination remains unclear. However, this study does not require CT examinations, which are necessary for measuring spleen volume, and can be done only by collecting blood, so it has the advantage of being easily tackled in future prospective studies.

In conclusion, splenomegaly, prominent thrombocytopenia, and liver dysfunction were confirmed after oxaliplatin-based adjuvant chemotherapy, possibly owing to SOS. We believe that increased HA after oxaliplatin-based chemotherapy is strongly associated with these outcomes and may predict SOS.

The Authors declare they have no financial or other conflicts of interest.

Takashi M and HT designed the study. TM, HT, YS, TN, RK, HN, AH, YF, SM, DK, YT, NN, TM, HF, and NU performed data acquisition, analysis, and interpretation. Takashi M prepared the manuscript. TM revised the paper critically. All Authors read and approved the final manuscript.

We thank Jane Charbonneau, DVM, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.