Personalised molecularly matched therapies for carcinomas of unknown primary

Bookmark and Share
Published: 21 Sep 2020
Views: 852
Dr Jacob Adashek - University of South Florida, Tampa, USA

Dr Jacob Adashek speaks to ecancer in an online interview for the virtual ESMO 2020 meeting about personalised molecularly matched therapies for carcinomas of unknown primary (CUP).

He explains that patients with CUP currently have a poor prognosis due to the rarity of the diagnosis, so the research aimed to take advantage of next-generation sequencing techniques to improve the outcomes of these patients.

Dr Adashek reports that their highly personalised approach resulted in improved outcomes and discusses the potential impact of these findings beyond the treatment of CUP.

ecancer's filming has been kindly supported by Amgen through the ecancer Global Foundation. ecancer is editorially independent and there is no influence over content.

Our study was personalised molecular matched therapies for carcinomas of unknown primary. Carcinoma of unknown primary is a rare tumour diagnosis that makes up about 2-5% of all cancer diagnosis worldwide and represents a heterogeneous group of cancers.

The unmet need of this study is that these patients fall into a category where they don’t have a specific tumour diagnosis so the standard of treatment includes combination chemotherapies with modest response rates between 20-40% and overall survival that’s poor with 6-15 months on average.

In this study we had 97 tumour samples with 74 of them got tissue next generation sequencing, 72 of them had cell-free DNA next generation sequencing. What we did with that was we found that the number of alterations in NGS on average the median was 4 with a range of 0-25 and the median for cell free DNA sequencing was 2 with a range of 0-9. We also had PD-L1 staining, microsatellite instability and tumour mutational burden in select cases.

From there we did what’s called the matching score where we looked at the number of targeted deleterious genomic alterations and divided that by the total number of deleterious alterations where we matched drugs to the target of the specific alteration.

What we found in this study was that the most commonly altered genes in tissue NGS were TP53 followed by CDKN2A and KRAS and in cell free DNA were TP53 followed by KRAS and PIK3CA. We found that about a third of patients had PD-L1 positivity where we had at least 1% on tumour staining and about 10% had high tumour mutation burden and about 3% had MSI high testing. From there we found out that patients with a high matching score defined as greater than 50% versus low matching score of less than or equal to 50%, including 25% of patients with no matched therapies, had a statistically significant longer median progression free survival of 10.4 months versus 2.8 months and a trend towards significance in overall survival of 15.8 versus 6.9 months.

We also calculated the disease control rate where we defined that as stable disease for greater than or equal to six months plus a partial or complete response and found that patients with high matching score versus low matching score had again a statistically significant higher disease control rate of 71% versus 24%.

The impact of these findings is that we have identified that patients with CUP have alterations that potentially could be targeted with targeted therapies. We also find that patients with CUP personalised combination approaches with high degrees of molecular matching or a matching score of greater than 50% are associated with improved outcomes.
So future directions include more prospective trials looking at matched targeted therapies including combination therapies to try and improve outcomes over standard of care for patients with CUP.

Are there implications from this study that can potentially be applied in other research?

Yes, I think the interest and the implications of this study is that in a cancer of unknown primary where you don’t have a specific tissue we’ve shown that when you target combination approaches and when you give multiple drugs to the patients they do better based on their NGS report. This can be translated or at least tried to be translated to other cancer entities and treat them more specifically based off of their mutational aberration patterns.