Oxaliplatin-Induced Sinusoidal Obstruction Syndrome Masquerading as Colorectal Liver Metastasis: A Case Report

Oxaliplatin-based chemotherapy is commonly used in adjuvant treatment of colon cancer as well as in neoadjuvant setting in patients with liver metastases. However oxaliplatin can cause damage to non-tumor bearing liver which presents as sinusoidal obstructive syndrome (SOS). These changes are difficult to differentiate from metastasis clinic-radiologically and manifests as sinusoidal dilatation, peliosis and nodular regenerative hyperplasia. The present study reports the case of a patient with oxaliplatin-induced SOS which mimicked colo-rectal liver metastasis on follow up imaging studies after receiving neoadjuvant oxaliplatin based chemotherapy. After multidisciplinary discussion, patient was planned for simultaneous resection of rectal primary and right hepatectomy for metastasis. Final histopathology revealed no tumour in liver but the liver lesions seen radiologically were actually changes of oxaliplatin induced focal SOS and mimicked metastatic nodules. In patients with colo-rectal cancer having received oxaliplatin-based chemotherapy, SOS may be considered as one of the causes of newly developed liver lesions, and should be subjected to additional radio-pathologic evaluation to prevent overtreatment and avoiding potentially morbid surgeries.


Introduction
Oxaliplatin, a third generation platinum derivative is widely used in management of many solid organ tumours. Oxaliplatin based chemotherapy regimens have been shown to improve overall as well as disease free survival in colorectal cancer [1,2]. Hence, Oxaliplatin based chemotherapy is commonly utilised in adjuvant and neoadjuvant settings in management of colorectal cancer as well as in management of colorectal liver metastasis. However, Oxaliplatin can lead to hepatocellular damage manifesting as Sinusoidal Obstruction Syndrome [SOS] which is a spectrum of pathologic changes ranging from sinusoidal dilatation to peliosis and nodular regenerative hyperplasia [3][4][5]. SOS is associated with increased intraoperative haemorrhage during liver resection and high post-operative morbidity. SOS can rarely present as an unusual conundrum with the SOS related liver changes mimicking liver metastases on radiologic imaging [6,7]. These are difficult to differentiate from metastasis clinico-radiologically and can be misinterpreted and over treated as metastatic disease. We hereby present the case of a 52-year-old gentleman with SOS where surgery was performed for suspicion of metastatic disease and final diagnosis of SOS was confirmed on pathology.

Case
A 52-year-old gentleman presented to our outpatient clinic with complaints of bleeding per rectum and loss of weight and appetite since 6 months. Magnetic resonance imaging (MRI) pelvis showed an upper-mid rectal growth involving anterior peritoneal reflection and enlarged mesorectal lymph nodes with extramucosal venous invasion and circumferential resection margin threatened at 9 o'clock. Computed tomography (CT) thorax and abdomen did not reveal any distant metastasis. Carcinoembryonic antigen (CEA) level was 47 ng/ml. Colonoscopy showed rectal growth at 5 cm from anal verge biopsy of which was reported as 1 3 well-differentiated adenocarcinoma. After a multidisciplinary meeting, he was planned for short course radiotherapy and neoadjuvant chemotherapy with clinical staging being cT4aN2M0. Patient completed short course radiotherapy (25 Gy/5#) and received 4# FOLFOX.
However, response dual phase CT scan of abdomen showed new onset hypodense liver lesions ( Fig. 1a and b) in segments V and VI of the liver with no contrast enhancement in porto-venous phase or delayed phase. Partial response at the primary site was also noted with anterior peritoneal reflection, mesorectal fascia and sphincter being free of tumour, changing stage to ycT2N1M1. CEA levels had decreased to 4.20 ng/ml. A targeted ultrasound examination of the liver was done by our experienced hepatobiliary radiologist who confirmed suspicious liver lesions. After multidisciplinary discussion, patient was planned for simultaneous resection of rectal primary and right hepatectomy for metastasis. Functional liver reserve for right hepatectomy was 39%. He was admitted and underwent middle hepatic vein sparing right hepatectomy along with laparoscopic low anterior resection and diversion ileostomy. Intraoperative ultrasound (USG) revealed 4 distinct lesions in segments V and VI of the liver with heterogeneous echogenicity without any other satellite lesions in left lobe. The segment VI lesion was palpable on surface and other lesions were identified on ultrasound examination. He had uneventful post-operative recovery and was discharged on postoperative day 6.
On gross examination, the liver parenchyma showed large brownish haemorrhagic area in the liver with areas of congested parenchyma in rest of the liver. No definite tumour nodule could be identified. Microscopy revealed large areas of haemorrhage and sinusoidal dilatation with focal increase in reticulin fibres. Mild distortion of liver architecture and mild chronic inflammation were also seen. This was probably attributable to oxaliplatin-induced changes (Figs. 2a, b and 3a, b) and no metastatic tumour was seen. The histopathology of primary tumour showed moderately differentiated adenocarcinoma with negative margins and reactive nodes (ypT3N0). Patient received adjuvant chemotherapy and follow-up CT scan after completion of treatment showed no residual or recurrent disease. Patient is alive and free of disease at the time of last follow-up 6 months after surgery.

Discussion
Oxaliplatin-based chemotherapy is routinely used as adjuvant therapy for colorectal cancer as well as in neoadjuvant setting in patients with liver metastases [8,9]. Sinusoidal obstructive syndrome (SOS), earlier known as hepatic venoocclusive disease, is a well-known side effect of oxaliplatin which manifests as sinusoidal dilatation, peliosis and nodular regenerative hyperplasia [3,4]. Deleve et al. [10] recognised that the disease process is initiated in the liver sinusoidal cells and renamed it as sinusoidal obstructive syndrome in 1999. The pathophysiology of SOS involves injury to sinusoidal endothelial cells by depolymerisation of F-actin and activation of matrix metallo-peptidase [MMP]-9 and MMP-2 leading to disintegration of sinusoidal wall [10][11][12]. The floating red blood cells are in turn extravasated into the space of Disse through the opened gaps (peliosis) and collagens get deposited at the space of Disse, exposed to the sinusoid (perisinusoidal fibrosis), leading to sinusoidal outlet obstruction. Though these changes are usually diffuse and involve whole liver, sometimes focal manifestations on non-tumour-bearing liver parenchyma are seen which can be misinterpreted as liver metastasis [6,13].
The incidence as well as imaging findings of these focally accentuated sinusoidal injury is not well-described in the literature and they are routinely interpreted to be metastatic lesions [13,14]. Ultrasonography shows heterogeneous echotexture with ill-defined lesions whereas on multiphase CT scan they are visualised as low attenuation/hypodense lesions with mostly no enhancement in porto-venous and delayed phases. Some cases may show heterogeneous enhancement [5]. On gadoxetic acid-enhanced MRI, these lesions are seen as hypo to iso-intense on pre-contrast phase with no enhancement on T1-weighted portal phase. They appear iso to slightly hyper intense on T2-weighted images [6]. Han et al. [6] compared the imaging features of pathologically proven chemotherapy-induced focal sinusoidal injury and that of metastatic liver lesion using liver MRI (using liverspecific contrast agents like Eovist). They reported ill-defined margin, non-spherical shape and intermingled signal intensity pattern in hepatobiliary phase to be specific findings of focal SOS. Ninety percent of the metastatic lesions had Fig. 3 a, b H&E stain, × 200, showing marked sinusoidal dilatation and congestion Fig. 2 a, b H&E stain, x 100 showing areas of haemorrhage in liver parenchyma peripheral rim enhancement on arterial and portal phases compared to these lesions. The non-tumour parenchyma was also seen to be less prominent on contrast imaging with lower apparent diffusion coefficient than normal liver but higher than metastatic masses. Presence of reticular hypo-intensities on hepatobiliary phase images of gadoxetic acid-enhanced MRI has been reported to be highly specific for SOS [6].
Qualitative imaging modalities like diffusion weighted MRI have superior diagnostic power than morphologic imaging modalities like CT or US in differentiating focal SOS and metastasis as diffusion restriction is seen in metastasis [13]. Positron emission tomography (PET) CT can also help in distinguishing metastasis from SOS as no 18-flurodeoxyglucose avidity is reported in focal SOS though less sensitive than MRI in diagnosis of liver metastatic lesions [14,15].
Various imaging modalities are available for diagnosis of colorectal liver metastasis. In the era of neoadjuvant chemotherapy, response assessment is dependent on accuracy of the imaging modality. Rojas Limpe et al. [15] compared various imaging modalities in resectable colorectal liver metastasis patients with sensitivity of MRI being better than CT (91% vs 82%) and PET CT (91% vs 60%) in all patients. They reported similar sensitivity (90%) with MRI in subgroup of patients previously treated with chemotherapy. Both CT (77%) and PET CT (48%) were less sensitive to diagnose liver metastasis in post chemotherapy patients. MRI has been shown to be the most sensitive tool to detect colorectal liver metastatic lesions in recent meta-analyses [16,17].
In addition to MRI, CEA is another important factor in determining tumour burden and tumour response to neoadjuvant therapy. CEA levels correlate with tumour response with decent accuracy. Qi-Wen Li reported post chemotherapy CEA levels of < 1.97 ng/ml as one of the most accurate predictors of pathologic completion rates with a sensitivity of 73.6%, a specificity of 75% and accuracy of 74.6% [18]. Post neoadjuvant CEA change is also shown to be a superior predictive factor than imaging response when overall survival is considered [19,20].
With liver resections becoming safer, preoperative biopsy or fine needle aspiration (FNA) of resectable colorectal liver metastases is largely abandoned as it has been reported to have high risk of needle track seeding and decreased overall survival in addition to bleeding risks [21,22]. In exceptional situations like cases with atypical imaging or diagnostic dilemmas, preoperative biopsy or FNA could be contemplated in multidisciplinary discussion. We do not routinely perform liver biopsy or FNA for resectable or potentially colorectal liver mestases.
Although oxaliplatin-induced SOS may not present with symptoms, it may be associated with increased morbidity and bleeding risk in the perioperative period [4,5]. The present study reports the case of a patient with neoadjuvant oxaliplatin-induced SOS which mimicked colorectal liver metastasis on imaging studies. There was partial response in primary disease with reduction in tumour marker levels which in the hindsight should have prompted additional investigations like Gadoxetic acid-enhanced MRI, diffusion weighted MRI or PET-CT. In patients with atypical features on radiology not suggestive of metastasis, a preoperative biopsy or intraoperative frozen section may be taken into consideration. Presence of multiple suspicious lesions in segment V and VI prompted us to perform a formal right hepatectomy for this patient. However, parenchyma-sparing resection could also be an option for these patients.
In patients with colorectal cancer having received oxaliplatinbased chemotherapy, SOS should be considered as a differential diagnosis for new onset liver lesions, particularly if they possess representative findings of oxaliplatin-induced SOS and the lesions do not demonstrate FDG uptake on PET-CT prompting further work up. Going ahead, further studies elucidating incidence and imaging characteristics of focal SOS are required.

Conclusion
Thorough clinical knowledge of imaging findings and thus identification of focal SOS may help in preventing overtreatment and avoiding potentially morbid surgeries.
Author Contribution All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Shraddha Patkar, Vipul Gupta and Amit Chopde. The first draft of the manuscript was written by Vipul Gupta and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data Availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Ethics Approval and Consent to Participate
The study is compliant with ethical standards. Written informed consent was obtained from the patient for the surgery.

Conflict of Interest
The authors declare no competing interests.