In the 100kGP cohort, IBCs were characterised by more aggressive tumour characteristics than SDBCs - a higher histological grade, larger tumour size, more frequently lymph node-positive at diagnosis, a higher proportion of ER and PR negativity, and more likely to be TN. These observations are consistent with several published studies7,38–55. While not statistically significant, in keeping with a recent study, IBC patients were more likely to be carriers of BRCA-mutations than those with SDBC (4.5% vs 2.5%)56.
Adjusting for grade, IBCs displayed a higher mutation rate while copy number abnormalities were not significantly different. IBCs additionally showed salient differences in driver mutational profile notably with a lower frequency of PIK3CA mutations and higher frequency of TP53 mutations. These TP53 mutations in IBCs were more likely to be clonal, supporting the notion that IBCs have differing evolutionary histories when compared to SDBCs11,21. The IBCs were also more likely to display dHR, which in part may reflect a higher rate of germline BRCA mutations. This is in keeping with published studies reporting a higher rate of germline HBOC predisposition in IBC compared to SDBC, as well as elevated rates of interval cancers in HBOC mutation carriers9,56–62. We did not, however, find evidence to indicate that women with IBC are at a lower PGS-defined breast cancer risk which has been reported by another analysis63. Similarly, while it has been proposed that IBC may be associated with a different immune response to SDBC64, our analysis of genetically mediated immune evasion or T-cell tumour infiltration provides no support for such an assertion.
While a number of our findings are consistent with a recently published study on IBCs from the Lifepool cohort9, there are a number of notable differences. In a smaller cohort of 46 IBCs and 64 SDBCs, except for dHR, no difference in somatic aberrations between IBCs and SDBCs was found. Additionally, Cheasley et al. report cancers in low mammographic density breasts (n = 142) had a higher frequency of TP53 mutations, dHR, higher fraction of the genome altered, more copy number gains and were more likely to be interval breast cancers when compared to cancers in high mammographic density breasts (n = 119)64. Whilst sample sizes and case ascertainment may explain this discrepancy, the IBCs in 100kGP appear to have molecular features common to the breast cancers in low mammographic density breasts in Lifepool. Since mammographic sensitivity for breast cancer detection increases with decreasing breast density, this provides further evidence that the breast cancers in the 100kGP represent rapidly growing breast cancers rather than breast cancers missed at screening due to technical factors65.
The major strengths of our study are having access to a unique dataset with high-quality WGS data, linkage to high quality clinicopathological information features on patients screened in the context of population-wide screening, thus avoiding biases related to screening indication. Furthermore, rather than focusing on a restricted set of genes we have been able to undertake a systematic analysis of the genetic landscape of IBCs. This has allowed us to assess tumour intrinsic mechanisms contributing to IBC as well as surrogates for the tumour microenvironment (TME). We do however, acknowledge a number of limitations, notably the lack of expression data and other classifiers of breast cancer histology such as the Gallen subtypes or PAM5066,67. Moreover, we made use of indirect measurements of the TME and are unable to assess the impact of immune dysfunction in the earliest stages of tumour development. Finally, in assessing the role of breast density and other modifiable factors we have relied on PGS, which inevitably affords limited power to demonstrate a relationship since these only capture a small proportion of the phenotypic variation of each risk factor33,68.
As well as informing on the biology of IBCs, the findings of our study may inform on the detection and management of breast cancer. The higher number of TP53 and copy number alterations as well as an increased mutation rate and dHR, suggests genomic instability has a greater role in IBCs when compared to SDBC. Genomic instability may be exploited therapeutically through synthetic lethality69,70, such as sensitivity to PARP in the context of BRCA-deficiency71,72. Furthermore, given 30% of breast cancers in women regularly attending screening are IBCs, additional screening modalities have been advocated as a means to advance the early detection of breast cancer. Whilst early in development, noninvasive detection of chromosomal instability in plasma circulating cell-free DNA has shown promise and may serve as an adjunct to breast screening mammography73,74.
In conclusion our findings indicate that while IBCs may not represent a distinct molecular subtype of breast cancer they are characterised by a more aggressive phenotype, in part likely to be a consequence of the timing of tumour initiation. As well as being relevant to informing patient management our findings provide a strong rationale for the UK to transition from three-yearly mammography to a program based on biannual screening, which has commonly been adopted by many countries75.