Thread by @CharlesSwanton

Excited to share our recent @SwantonLab @NickyMcGranahan & @MariamJHanjani lab work on the genomic evolution of non-small cell lung cancer metastases in #TRACERx, out @Nature

nature.com/articles/s4158…

A quick thread [1/13]

We analysed paired primary-metastasis data from 126 patients (218 metastatic & 476 primary tumour samples) within the first 421 patients recruited to TRACERx. We explored the timing of metastatic divergence, modes of metastatic dissemination and selection in seeding clones [2/13]

We reconstructed primary tumour phylogenies & timed metastatic divergence relative to the trunk. Divergence is when the metastatic clone first existed & differs from time of spread. The majority of cases (94/126) diverged late following a clonal sweep in the primary tumour [3/13]

Early divergence (32/126) was associated with being a smoker at the time of primary tumour resection, suggesting that smoking may provide fuel for ongoing clonal sweeps in the primary tumour [4/13]

But what does this mean in terms of tumour sizes? Simulations suggest for early divergence cases, the seeding clones likely existed when the tumours were <8mm, the limit at which many solid tumour nodule screening programmes recommend further investigations [5/13]

This suggests that screening should work in the majority of cases, but in a minority, tumours may be aggressive before we have the chance to detect them using screening programmes [6/13]

Can we time divergence in other ways? Yes! For example, we used the platinum mutational signature to time the divergence of 2 brain metastases to a period when adjuvant therapy was given: the signature was detected in the occipital & not the cerebellar metastasis [7/13]

Next, we explored the dissemination patterns of metastases from the primary tumour. We were able to determine whether metastatic seeding involved a single clone (monoclonal) or whether multiple clones were involved in seeding (polyclonal) [8/13]

We found that the vast majority of cases (68%) exhibited monoclonal dissemination from the primary. Polyclonal dissemination was associated with extrathoracic disease relapse. Undersampling of metastases has likely underestimated the frequency of polyclonal dissemination [9/13]

But do metastases then seed other sites? Yes they do! Interestingly, lymph nodes resected with the primary tumour seed <20% of recurrences/progressive disease, suggesting that they are usually a hallmark of metastatic potential rather than a gateway to metastases [10/13]

Are there any characteristics that distinguish seeding & non-seeding clones? We find that seeding clones are usually fitter in the primary tumour as evidenced by having a higher cancer cell fraction & occupying larger areas in the primary tumour than non-seeding clones [11/13]

This was a massive group effort by the #TRACERx consortium. We would like to thank all the co-authors @drmaiseb @AMIHuebner @CarlosEcolEvol @KristianaGrigo4 @tbkwatkins @oriol_pich, supervisors @MariamJHanjani @NickyMcGranahan, & Nature editorial team @mvicaracal @Nature [12/13]

Finally, we would like to thank all the patients and their families for taking part in this study, and our funders.
@royalsociety @CRUKLungCentre @CR_UK @CRUKresearch @CRUKCOLcentre @RosetreesT @uclcancer @UCLHresearch @NIHRresearch @BCRFcure @TheCrick @wellcometrust
[13/13]