Preoperative Chemotherapy on Functional Liver Regeneration for Colorectal Liver Metastases Assessed With ^99mTc-GSA SPECT/CT Imaging

From January 2009 to December 2013, we enrolled 40 patients who underwent major hepatectomy and functional assessment of the future remnant liver at Kumamoto University Hospital's Department of Gastroenterological Surgery. Major hepatectomy was defined as the resection of ≥3 liver segments, according to the International Hepato-Pancreato-Biliary Association.²⁰  Low-grade liver fibrosis was defined as F0 or F1 of pathologic fibrosis staging in the Inuyama classification.²¹  Viral hepatitis was defined as positive test results for the hepatitis B surface antigen or anti-hepatitis C virus antibody. All procedures involving human participants were conducted in accordance with the ethical standards of the institutional and/or national research committee and the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. The Institutional Review Board of the Graduate School of Life Sciences, Kumamoto University, approved this study (approval number: 717; December 25, 2007). Informed consent was obtained from all the participants in the study.

According to our previous reports,^11,22–24  we applied a SPECT/CT system consisting of a dual-head SPECT camera and a 16-row multisection CT scanner (Symbia T16, Siemens Healthcare, Erlangen, Germany) for scintigraphy and CT imaging, respectively. After the completion of dynamic scintigrams and static SPECT images, contrast-enhanced CT was obtained. Briefly, SPECT and CT images were fused using a dedicated workstation (Ziostation2, ZIOSOFT, Tokyo, Japan). The total liver volumes, excluding tumor and future remnant liver volumes, were determined using the contrast-enhanced CT images. Percent remnant liver volume (%LV) was expressed as a percentage of the total liver volume. From the ^99mTc-GSA SPECT/CT fused images, we calculated liver uptake value (LUV) as a quantitative liver functional index^11,23  for the whole liver and future remnant liver. The percent remnant functional liver volume (%FLV) was calculated as LUV of the future remnant liver/ LUV of the whole liver × 100. The ^99mTc-GSA SPECT/CT fused images were obtained within 2 weeks prior to liver resection and at 1 month later.

The type of hepatectomy was selected based on the size, number, and location of the tumor; extent of vascular invasion; %LV; %FLV; and the patients' age or general condition, as described previously.^19,25  To reduce the intraoperative blood loss, we generally used the precoagulation technique, Pringle's maneuver, or hanging maneuver.²⁶  Laparoscopic liver resections were accepted for the patients with CRLM and the other tumors.^27,28 

Values were expressed as median (range) or mean ± standard deviation as required. Median and mean values were estimated using the Wilcoxon or Kruskal–Wallis test, Pearson test, and Student's t-test. For statistical analysis, we used a statistical software package (JMP, version 9, SAS Institute, Cary, North Carolina). A value of P <0.05 was considered statistically significant.

A total of 40 patients underwent major resection and were divided into 2 groups: CRLM with preoperative chemotherapy (CT-CRLM; n = 12) and a control group without preoperative chemotherapy (n = 28). All the CT-CRLM patients who underwent major hepatectomy exhibited low-grade fibrosis (F0 or F1 according to the Inuyama criteria)²¹  in the background liver and no viral hepatitis. Chemotherapy-naïve CRLM patients were the best candidates as a control group; however, the number of major hepatectomies among patients who received upfront hepatectomy for CRLM was small. Therefore, as a control group, we selected 28 patients from our database who underwent major hepatectomy without preoperative chemotherapy and did not have severe fibrosis and viral hepatitis. The pathologic diagnoses were CRLM (n = 15); hepatocellular carcinoma (n = 9); intrahepatic cholangiocarcinoma (n = 7); bile duct carcinoma (n = 4); and other tumors (n = 5). Patient characteristics are summarized in Table 1. All the background factors were identical, with the exception of total bilirubin levels and platelet counts; however, the median values were within the normal range for both the groups.

Serial changes in the volume and functional volume of the preoperative total liver, future remnant liver, and post 1-month liver measurements are displayed in Figs. 1a and 1b, respectively. After 1 month, %LV was similar (P = 0.966) in the CT-CRLM and control groups (74.7% ± 13.4% and 74.5% ± 11.2%, respectively). Similarly, %FLV after 1 month was similar (P = 0.565) in the CT-CRLM and control groups (98.1% ± 10.7% and 98.7% ± 10.7%, respectively). Furthermore, after 1 month, %FLV was significantly larger than %LV in the CT-CRLM and control groups (both P < 0.001).

In the CT-CRLM and control groups, the percentage increases in LV were 39.3% ± 29.0% and 23.2% ± 23.5%, and FLV were 79.4% ± 43.1% and 57.0% ± 33.4%, respectively (Figs. 2a and 2b). The percentage increase in LV was significantly higher in the CT-CRLM group than in the control group (P = 0.037), but the percentage increase in FLV was similar in the 2 groups (P = 0.417). The absolute differences in LV were 216.2 ± 155.7 cm³ and 148.7 ± 134.7 cm³ and FLV were 19.4 ± 8.5 %/m² and 17.4 ± 7.9%/m², respectively (Figs. 3a and 3b). The absolute differences in LV and FLV were equivalent between the 2 groups (P = 0.086 and 0.235, respectively).

Median postoperative total bilirubin levels (0.6 mg/dL versus 0.7 mg/dL); prothrombin activities (93% versus 93%); the uptake ratio of the heart at 15 minutes compared to 3 minutes (0.58 versus 0.53); and the uptake ratio of the liver compared to the liver plus the heart at 15 minutes (0.92 versus 0.94) determined through ^99mTc-GSA scintigraphy were equivalent between the CT-CRLM and control groups.

The data of 15 CRLM patients are described in Table 2. The median age was 64.7 ± 11.6 years, and the female-to-male ratio was 6:9. In 12 CRLM patients with preoperative chemotherapy and 3 chemotherapy-naïve CRLM patients, the mean resection rate/functional resection rate was 44.3% ± 12.3%/42.3% ± 11.5% and 42.7% ± 11.1%/39.9% ± 7.3%, and the mean percentage increase in LV/FLV from the baseline values of the future remnant liver was 36.6% ± 27.7%/73.4% ± 37.7% and 24.5% ± 29.0%/145.4% ± 55.1%, respectively. There were no significant differences between CRLM patients with and without chemotherapy. Among the 12 CRLM patients with preoperative chemotherapy, OX; irinotecan (IRI); and OX + IRI were administered to 8, 1, and 3 patients, respectively. Bevacizumab was initiated in 3 patients, antiepidermal growth factor receptor antibody in 2 patients, and both treatments were administered to 3 patients. The median number of preoperative chemotherapy courses was 8 (range: 6–16). The number in treatment of 3-week regimens was counted as 1.5 times of 2-week regimens. Various types of major hepatectomies were performed. We observed no distinct differences in LV and FLV with regard to the regimens and cycles of preoperative therapy. We assessed 6 out of 15 (40%) patients as having grade 2 or greater sinusoidal obstruction⁹ ; all the patients received OX-containing chemotherapy, and 4 patients were treated without bevacizumab. No patients exhibited steatohepatitis with nonalcoholic fatty liver disease activity score >4.²⁹ 

Recently, we have reported that the functional liver regeneration, and not the volumetric liver regeneration, as assessed by ^99mTc-GSA SPECT/CT fused images, can correctly predict the levels of postoperative liver dysfunction.¹¹  The functional liver regeneration was influenced by the degree of liver fibrosis and the functional resected volume, and the existence of viral hepatitis.^12–14,16  Therefore, as a control group, we selected patients who underwent major hepatectomy without preoperative chemotherapy and who did not have severe fibrosis and viral hepatitis. The preoperative background factors, including liver function, functional resection rate, and future remnant FLV, were almost equivalent. Total bilirubin levels and platelet counts significantly differed between the 2 groups; however, the median values were within the normal range. Because platelet-derived serotonin is associated with the initiation of liver regeneration, decrease in platelet counts may negatively impact liver regeneration.¹⁷  In the present study, both liver regeneration and functional liver regeneration parameters were equivalent between the CT-CRLM and control groups. Conversely, percentage increases in LV were somewhat greater in the CT-CRLM group.

Modern chemotherapy for CRLM can induce various liver parenchymal injuries such as steatosis, steatohepatitis, and sinusoidal obstruction. Liver regeneration within 1 week after hepatectomy did not differ between patients with and without steatosis.³⁰  However, when only patients with moderate-to-severe steatosis (steatosis ≥30%) were considered, the late liver regeneration after 6 months in the steatosis patients was lower. In a mouse model, steatosis did not impair liver regeneration after partial hepatectomy.³¹  Moreover, simple steatosis prompted by Western diet–induced steatosis can enhance liver cell proliferation, which was accompanied by increased hepatocyte growth factor and leptin signaling. In contrast, in rats fed on a standard methionine- and choline-deficient diet, steatohepatitis probably impaired liver regeneration because of increased hepatocellular lipid peroxidation and damage in concert with Kupffer cell–mediated inflammatory responses.³²  In the present study, no patients exhibited moderate-to-severe steatosis (steatosis ≥30%) or apparent steatohepatitis.

VEGF is one of the key molecules for liver regeneration, and bevacizumab is an anti-VEGF monoclonal antibody. In the present study, 6 patients received bevacizumab-containing chemotherapy; however, no obvious liver regeneration failure was noted. Meanwhile, OX-based chemotherapy can induce sinusoidal obstruction, with the risk of impairing postoperative liver regeneration,^5,6  and this adverse effect may be avoided by using bevacizumab but not cetuximab or panitumumab.^8,9  A case-matched study recently demonstrated that bevacizumab did not impair liver regeneration after major hepatectomy even in patients with high exposure to preoperative chemotherapy.³³  In a clinical report, liver regeneration was not adversely affected by chemotherapy, including molecular-targeted drugs for patients with a high grade of steatosis or sinusoidal obstruction in the surrounding liver parenchyma.³⁴  Portal vein embolization (PVE) can provide liver regeneration similar to that by major hepatectomy.²²  One report described that chemotherapy with bevacizumab did not affect liver regeneration after PVE.³⁵  In contrast, bevacizumab may impair liver regeneration after PVE, and an age >60 years and treatment with ≥6 cycles of bevacizumab were risk factors for poor regeneration.³⁶  In the present study, 6 out of 15 (40%) patients were assessed as having grade 2 or higher sinusoidal obstruction; all patients received OX and 4 were treated without bevacizumab. In all 15 CRLM patients, FLV increased, with a mean increment of 73.4% ± 37.7%; conversely, LV decreased in 2 patients, and the mean increment was 36.6% ± 27.7%. The percent increases in LV and FLV were not always correlated patient-by-patient (data not shown). There were no obvious differences in both volumetric and functional liver regeneration with regard to the differences in chemotherapeutic regimens.

To our knowledge, the present report is the first to elucidate that both functional liver regeneration and volumetric liver regeneration after major hepatic resection are unaffected by preoperative chemotherapy for CT-CRLM patients; however, our study has some limitations. First, for CRLM, limited hepatectomy has been more frequently selected instead of major hepatectomy, and preoperative chemotherapy has often been employed^1,2 ; therefore, the number of patients undergoing major hepatectomy, particularly without preoperative chemotherapy, was small. In this paper, the results of the CT-CRLM patients were mainly compared with those of patients with various liver tumors. However, there were no significant differences in liver regeneration data between CRLM patients with and those without chemotherapy. Second, various chemotherapies were administered in a neoadjuvant or conversion manner^2–4 ; therefore, the number or duration of chemotherapy was not constant. Third, our study did not include cases of extremely extensive resection or resection in extremely damaged livers. Finally, the assessment of functional liver regeneration with ^99mTc-GSA SPECT/CT fused images has not been applied outside Japan.

Preoperative chemotherapy may not affect functional liver regeneration in patients who receive several courses of OX- or IRI-based chemotherapy with molecular-targeted drugs. Further prospective and large-scale studies are required to confirm these results, using the same regimen and treatment duration in CRLM patients with or without preoperative chemotherapy.

All authors have no conflict-of-interest about this study, and have not received any financial or material support.