Patient-Derived Organoids Show Promise for Personalized Therapy in Pancreatic Cancer

Manuel Hidalgo Medina, MD, PhD

Patient-derived organoids (PDOs) may possess the potential to personalize therapy and allow for the exclusion of ineffective drugs from combination regimens, reducing toxicity in patients with pancreatic cancer, according to preliminary results from the HOPE trial.¹ As such, it is anticipated that this approach will be explored in future clinical trials to prospectively inform treatment selection for this patient population, according to Manuel Hidalgo Medina, MD, PhD.

In the trial, investigators generated PDOs from patients with pancreatic cancer under real-world conditions to evaluate drug sensitivity against PDOs and correlate these data with clinical outcomes. At the data cutoff, which was January 2020, a total of 76 patients with all stages of the disease were enrolled on the trial. Drug testing was successfully performed on PDOs generated from 12 of these subjects. The success of obtaining sufficient cells for generation depended on factors such as modality, body part, and tumor cellularity.

Investigators estimated that a PDO AUC value of <1.66 yields a >99% probability of disease control from a regimen that contains a certain drug. They also found that if all drugs in a regimen had an AUC of >2.75 there is a >80% probability of accurately predicting resistance.

In a patient with stage IV pancreatic ductal adenocarcinoma (PDAC) with a KRAS mutation and ERBB2 amplification, the investigators were able to demonstrate the potential of PDO testing to tailor treatment. The patient experienced disease control with FOLFIRINOX, which was held for toxicity. The PDO demonstrated resistance to oxaliplatin and subsequently, the patient experienced an extended period of disease control with regimens that did not include the drug. These findings demonstrated the potential of PDO drug sensitivity testing to exclude ineffective treatments from combination therapy and thus, limit toxicity.

"We think that the correlation [between drug sensitivity and PDOs] is there, and now we need to industrialize the development of the organoid and standardize it for it to be a clinically applicable tool," said Hidalgo, who is lead author of the study. "That will be a part of a subsequent clinical trial."

Additionally, Hidalgo presented results from a phase 1b study which examined the use of palbociclib (Ibrance) plus nab-paclitaxel (Abraxane) in patients with metastatic adenocarcinoma of the pancreas. Results showed that the objective response and disease control rates in the 23 patients who were treated with the MTD were 13% and 65.2%, respectively. In these patients, the median progression-free survival, median overall survival, and survival probability at 12 months were 5.3 months (95% CI, 3.5-9.7), 12.1 months (95% CI, 6.4-14.8), and 50% (95% CI, 29.9%-67.2%), respectively.²

"The take-home message is that we think there is potential here, but we need to do more preclinical work to properly design the next generation of clinical trials," said Hidaglo. "As it stands, I would not do a phase 3 trial based on [the current findings]."

In an interview with OncLive, Hidalgo, chief of the Division of Hematology and Medical Oncology of Weill Cornell Medicine and NewYork-Presbyterian Hospital, discussed some of the research he presented at the 2020 ACCR Virtual Annual Meeting and future areas of exciting efforts being made in pancreatic cancer.

OncLive: What is the rationale for generating PDOs in pancreatic cancer?

Hidalgo: PDOs are a very common tool these days; they have been generated by different methodologies, from many subsets of patients and diseases. We have been interested in [their use in] pancreatic cancer for some time. The goal of the HOPE trial was to use PDOs to predict treatment response [in this patient population]. We wanted to develop organoids as a tool to do that. In this first trial, we basically correlated response of patients to a standard of care with the response of the organoid to the same standard of care. We classified patients as responders and non-responders and we tried to correlate that with a response in the organoid. In the subset of patients [from whom we] were able to [generate] the organoids and test them, the correlation seemed to be quite positive. That has been reported by other groups, as well.

How are PDOs being used in this space in general?

PDOs are being used for many different strategies. One is to test drugs and prioritize which drugs will be more effective in a particular disease. PDOs are a good screening tool for drugs.

We are personally more interested in the precision medicine angle. We want to build collections of organoids, test them against different drugs, and use that information to treat patients. It is also a tool for biomarker development. As you analyze the organoid, you can learn more about genomics, epigenetics, and proteomics. Then, you determine which drugs the organoid is susceptible or resistant to and you can get information as to potential biomarkers.

What inspired the HOPE trial?

The research was inspired by our interest in developing approaches that will help patients. The goal of this pilot feasibility trial was to test whether we could do this, whether it could be done in real-time, whether the approach was feasible, and what kind of biopsies we would need to get. We also wanted to find out whether there is a correlation strong enough to build a hypothesis on this and then design a second trial, HOPE-2, that will more formally test the hypothesis. It’s safe to say we got what we needed.

Could you walk me through the methods used?

You take a biopsy, which can be a needle biopsy or a surgical specimen, and then you take that specimen to the lab, desegregate cells, and plate the cells in the matrigel and organoid-containing medium. We then develop what the organoid needs to take in the plate in order to grow. Then, when we have enough material, we start to test drugs and measure cell death.

What were the preliminary findings presented at the meeting? Were any of them particularly surprising?

What was surprising was that we learned that it is not easy to get the organoids up and running in real clinical time. We missed many patients either because their biopsies were not big enough, we did not process the biopsies rapidly, or because we did not plate them correctly. As such, it has been a learning exercise on one end.

The positive side is that when a drug works nicely in the organoid, and we when went back and looked at the patient who had been treated with that drug, the correlation was pretty good.

Do any challenges exist with using this kind of approach?

The biggest challenge has to do with getting all the logistics together.

Is this approach financially feasible?

If we can personalize drugs for patients and avoid utilizing drugs that are not useful, it will be. You spend on one side, but you are saving a lot on the other side [when you are able to avoid] drugs that do not work, that create toxicities, and [lead to increased hospital] admissions. That is very helpful.

Do you see PDOs being used more often as a way to personalize medicine?

Totally. It is going to be a very intense area of research that will be done by many groups. Papers are also constantly coming out on this topic.

Are any future trials planned that will be examine PDOs further?

We are going to design the HOPE-2 trial, which will be a proper phase 2 trial that will test whether or not [this approach] really works.

What is the key takeaway from this poster that you want your colleagues to know?

With tumors like pancreatic cancer, where we do not find actionable mutations, looking at phenotypic screening can be another way to personalize treatment for our patients.

Moving on to your other presentation that was given at AACR, on a phase 1b study with palbociclib plus nab-paclitaxel in patients with metastatic adenocarcinoma of the pancreas. What was the rationale for this research?

Several published papers have showed that CDK4/6 inhibitors may play a role in pancreatic cancer. Initially, we thought that because those tumors often have deletions or mutations in B16 and, therefore, they have activation of CDK4/6. For that reason, [it has been hypothesized] that palbociclib could be a chemical B16 and will block that, and therefore block the cell cycle. As such, we developed that preclinically and put it in combination with chemotherapy. We saw in the preclinical studies that the combination with taxanes was as effective as the combination with gemcitabine and taxanes. Therefore, we could evaluate a doublet rather than a triplet regimen. This trial explored different sequencings and doses and ways to administer these drugs.

The results were reasonable, but we did not meet the prespecified criteria for survival. It was pretty ambitious, but that regimen will not move forward. We learned a lot in preclinical studies though, and I think we will come back to [re-approach this concept] in a different and, hopefully, more effective manner.

Could you expand on the findings?

We found that the combination is tolerable and that the adverse events are within the expected range. The responses and the survival observed with the combination was good, at 50% at 12 months, but we wanted to see a 65% rate. No drug interactions were observed.

What are some of the questions you still want to answer regarding this combination?

We believe that a role for combining chemotherapy with CDK4/6 inhibitors exists—not concomitantly, but sequentially. We believe that if you give chemotherapy first, followed by CDK4/6 when cells are recovering from chemotherapy insult, you can achieve a better response.

Are there any other research efforts that you're personally involved in, that we didn't discuss?

We are working on [the development of several] new agents in pancreatic cancer. Personally, I have been working intensively with BL-8040, which is a CXCR4 inhibitor. Some data with this agent were presented at the 2019 ESMO Congress. Hopefully, we will be presenting and publishing new data this year, as well. That is one agent to keep an eye out for.

We have also been working on the CDK4/6 story to build upon it and continue developing organoids for personalized treatment. Additionally, [we have been] working with some interesting compounds. One is from Erytech Pharma, it is [called eryaspase] and it is in a phase 3 trial. We saw very interesting phase 2 data on it so we will see more when that study is complete.

I am working very intensively with the Pancreatic Cancer Action Network's Precision Promise platform for clinical research. A ton of clinical data are going to be coming from that group. These are exciting times, but we need to continue working on this. More progress is needed.

References

1. Grossman JE, Huang L, Muthuswamy L, et al. Organoid sensitivity in pancreatic cancer correlates with clinical response to treatment and reveals utility for reducing toxicity: Preliminary results from the HOPE trial. Presented at: the 2020 AACR Virtual Annual Meeting; April 27-28, 2020. Abstract CT119. bit.ly/3aMJ08Q.
2. Hidalgo M, Carbonero RG, Lim KH, et al. A phase 1b study of palbociclib + nab-paclitaxel in patients (pts) with metastatic adenocarcinoma of the pancreas (PDAC). Presented at: the 2020 AACR Virtual Annual Meeting; April 27-28, 2020. Abstract CT135. bit.ly/35ms0oF.