The development of the organoid model system has been a significant technological breakthrough of the past decade¹. Organoids offer a clinically relevant approach for understanding disease biology ranging from basic mechanisms of pathophysiology to treatment response².

Our laboratory at the Department of Biomedicine of the University of Basel, has a strong focus on the development of pre-clinical model systems derived from patients with primary liver cancer, namely hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (CCA). For the past 10 years we have mainly worked on the generation of patient-derived organoids (PDO)³ and patient-derived xenografts (PDX)⁴derived from diagnostic tumor needle biopsies, that we regularly obtain in our outpatient clinic from patients that provide written informed consent. The resulting biobank encompasses a collection of models that reflects the heterogeneity of tumors observed in patients and is a unique research resource for translational and personalized medicine.

HCC is an aggressive and difficult to treat disease mainly due to its pronounced heterogeneity at the clinical, histopathological and molecular level⁵. Understanding the factors that influence the response to a given treatment as well as the development of novel therapeutic approaches for patients with HCC are at the core of our research program. In the context of a routine diagnostic workup, we came across a patient with a lesion in the liver highly suspicious for HCC. A biopsy of the tumor and subsequent histopathological evaluation confirmed the diagnosis of HCC. The patient was eligible for surgical tumor resection due to the absence of cirrhosis and vascular invasion, and a favorable general condition. Comprehensive immunophenotypic characterization of the resected tumor revealed that the HCC also expressed markers of neuroendocrine differentiation. This particular tumor, termed hepatocellular carcinoma with neuroendocrine differentiation (HCC-NED), represents a very rare entity with only few cases described to date worldwide⁶. For this reason, no guidelines describing treatment recommendations in this rare tumor subtype are available^7-9. So far, these patients have mostly been treated with therapies approved for advanced HCC or neuroendocrine neoplasms.

A recognized feature of neuroendocrine liver cancers is their poor prognosis^7,10. In spite of surgical tumor resection in curative treatment intent, the patient had multifocal tumor relapse after only 12 weeks. Because of the tumor’s neuroendocrine differentiation, treatment choice was not clearly defined in any guidelines. For this purpose we used the tumor organoids that we routinely generated from the patient’s tumor (in the context of our biobanking project mentioned above). Using these organoids, we then performed an in vitroscreen of a set of compounds approved for HCC or neuroendocrine tumors and carcinomas. Using this approach, we hoped to identify compounds with preferential activity in organoids from HCC-NED as compared to a set of HCC organoids from our biobank. The immediately performed drug screening revealed that the combination of etoposide and carboplatin led to a better response in HCC-NED versus HCC organoids. Based on these results and considering the lack of a standardized treatment procedure for this tumor entity, the patient was treated with the organoid-informed combination therapy. However, after only two cycles of chemotherapy, the CT scan revealed progressive disease.

We therefore went back to the genomic tumor data to investigate, whether the HCC-NED tumor harbored any potentially druggable mutation. Besides TP53 and CTNNB1 hotspot mutations, the tumor displayed a NTRK1 mutation of unknown significance (without TRK overexpression). As pan-TRK inhibitors have recently been described to strongly inhibit growth of gastroenteropancreatic neoplasms¹¹, we also investigated the antitumoral activity of entrectinib and larotrectinib in the HCC-NED organoids. Larotrectinib displayed no activity, however entrectinib lead to growth inhibition. Taken together, these results suggest that the NTRK1 mutation present in this tumor was not a driver mutation; the effect of entrectinib was rather due to its additional activity against ROS1 and ALK¹².  

Based on these results, the patient obtained entrectinib as a second line therapy. However, the patient’s general condition deteriorated rapidly and entrectinib treatment had to be discontinued after only two weeks (therefore formal evaluation of treatment response was not possible). One week later the patient unfortunately deceased.

In our liver cancer study cohort, this was the first patient where personalized treatment choice was possible based on the results obtained in our ex vivo drug screening using PDOs. On the one hand, this was possible due to the short duration time until establishment of the organoid line, which is probably a consequence of the tumor’s high proliferation rate and therefore its clinical aggressiveness. On the other hand, the patient first obtained surgical tumor resection in curative intent, therefore not requiring immediate systemic treatment.

In our case, combination treatment with etoposide and carboplatin was not effective in the patient despite its antitumoral activity in the organoids. We can only speculate on the reasons for this discrepancy. First, the organoid model lacks the tumor microenvironment (especially stromal and immune cells) which might affect treatment response. Second, the tumor might have developed rapid resistance in vivo. Furthermore, drug concentrations achieved in vivo might have been insufficient or treatment readouts might not be comparable (radiological evaluation based on the mRECIST criteria in vivo vs. ATP/proliferation-based readout in vitro).

References