Intrinsic subtypes in Ethiopian breast cancer patient

In this prospective cohort of Ethiopian patients with BC, intrinsic PAM50-based subtypes of tumors were determined in addition to receptor status by immunohistochemistry (n = 334). Both methods had considerable similarities, especially for basal-like or triple-negative and HER2-enriched or HER2-positive types. The high proportion of patients below the age of 50 (approx. 60%) and large tumor sizes (T3 and T4, approx. 40%) differed from high-income settings. This probably resulted in the lower proportions of luminal A subtypes compared to Western cohorts.

Pattern of subtypes

We reported in 2014 that two thirds of tumors from a patient cohorts from Addis Ababa had endocrine responsive disease [13]. Another cohort from rural Ethiopia showed more than half of the patients with positive hormone receptor status [6]. The third study published in 2018 again from Addis Ababa also reported 65% of receptor-positive disease [21]. Within the current study, immunohistochemistry as well as RNA-expression analysis were used to assess endocrine responsiveness. The results also revealed that even using RNA-expression based subtyping, more than half of the tumors were endocrine responsive. In detail, we found 31.1% luminal A, 27.2% luminal B, 18.6% HER2-enriched and 23.1% basal-like tumors. This is considerably shifted to the more aggressive subtypes compared, for example, to data from the United States (Nurses Health Study) that reported 46% luminal A, 18% luminal B, 14% HER2-enriched, 15% basal-like, and 8% normal-like subtypes [22]. Such direct comparison needs to be interpreted with care.

Comparing with other ethnic groups

High quality data on tumor subtypes from the cancer genome atlas about tumors from African patients is lacking. Patients with African compared to European ancestry had a higher likelihood of basal-like (odds ratio OR 1.67) and HER2-enriched (OR 2.22) tumors [23]. African–American patients genetic background has been reported as predominantly from West Africa [24].

Geographic variations have been observed across Africa concerning the composition of different hormone receptor-based types no longer characterizing all tumors in Africa as having aggressive phenotypes. A large meta-analysis found that patterns of tumor subtypes in East Africa appear to be more favorable compared to other geographic regions such as West Africa [12].

Other factors influencing subtype composition

Our convenient hospital cohort has specific features. The age composition as well as proportions of early and later stage disease may influence the overall proportions of subtypes. In general, cohorts from sub Saharan Africa seem to similarly represent a young population and commonly comprise of late stage tumors compared to Western settings. It was shown in a cohort from the United States that patients with luminal A tumors are on average 5–6 years older than patients with the other subtypes [25]. A comparison of receptor status between Sudanese and German patients also showed considerably more low-grade, receptor-positive tumors in older German women [26]. Therefore cohorts with fewer older patients lack these luminal tumors. Additionally, body composition has been reported to influence subtype. Obesity has been associated with higher risk of luminal A type breast cancer [27]. Since obesity is still rare in Ethiopia, less luminal A breast cancers are expected [28].

Comparing subtypes and IHC

Assessing tumor biology with RNA-expression analysis is considered gold standard, at the same time immunohistochemistry has been used as basis to prove effectiveness of endocrine treatment. When comparing both methods, certain discrepancies can be seen. Consistent with previous reports [29], we have observed a substantial mismatch between the classification based on gene expression (PAM50 intrinsic subtypes) and on immunohistochemistry (IHC groups using ER, PgR and HER2). Still, hormone receptor-positive tumors were mainly classified as luminal subtypes. Therefore, immunohistochemistry results appear reliable as a basis for decision to advise endocrine treatment to the patient despite lack of standardized external quality assurance measures for tissue processing and immunohistochemistry performance [30].

In Ethiopia, as well as many other countries in Sub-Saharan Africa, basic immunohistochemistry is not always accessible to all patients. Even in capital cities, 15 out of 20 centers experienced frequent power cuts and four out of twenty had no immunohistochemistry in the country [31]. Due to the fact that also for Ethiopian patients’ routine immunohistochemistry is not always available, several confirmatory studies reported that half of patients have endocrine responsive disease (confirmed by RNA-expression as well as immunohistochemistry studies). This may provide clinicians more confidence in their decision process to offer endocrine treatment for patients with unknown receptor status.

Individualized therapy

As a future perspective, multiplexed hybridization assay-based subtype determination via nCounter® can also support individualized therapy in Ethiopia. Since a large proportion of tumors are larger than 2 cm or lymph node-positive, without additional prognostic markers, according to the National Cancer Control Network (NCCN) harmonized guidelines, nearly all patients would need chemotherapy. As an example, utilizing PAM50-based subtypes, we were able to split the large group of patients with HR-positive and HER2-negative tumors into clinically relevant, more homogenous intrinsic subtypes. Patients with HR+/HER2 tumors have been regarded as an especially challenging BC group due to high variance concerning clinical outcome [32]. These patients include cases of low, intermediate and high risk of recurrence. Differentiating the HR+/HER2 group into luminal A and non-luminal A intrinsic subtypes increases the ability to assess recurrence risk in a more individualized manner, since chemotherapy is not recommended for patients with luminal A tumors [33, 34]. This information also enables clinicians to counsel patients on the importance of adherence to adjuvant endocrine therapy long-term and surveillance for local or distant recurrence.

Individualized therapy: HER2-enriched and basal-like types

The subtype determination can also personalize treatment for patients with aggressive tumors. The PAM50 assay allowed the identification of 62 tumors as HER2-enriched, regardless of HER2 status. Using the IHC method, only 41 patients would get an anti-HER2 therapy. This means that the PAM50 assay leads to an approximately 50% incremented number of patients who would benefit from drugs suppressing the HER2 signaling pathway (anti-HER2 therapy). Our results are concordant with the seminal work of Perou and colleagues in the sense that two thirds of our samples classified as HER2-positive were found to be HER2-enriched (Table 1) [10].

Finally, 77 samples were classified as basal-like through the PAM50 assay. Out of these, only 46 (60%) were concordant with the triple-negative class via the IHC method. These samples are arguably the ones that would benefit most from chemotherapy, 40% of the basal-like samples are non-TNBC [10]. Conversely, approximately 38% of the TNBC tumors were non-basal-like. Thus, both PAM50 and IHC classification methods yield heterogeneous groups where personalized recommendations would improve by additional molecular subtyping as proposed before [35, 36].

Strengths and limitations

The strength of the study includes multi-center involvement in Ethiopia and the application of multigene assays for gene expression profiling which is the gold standard for molecular classification as well as hormone receptor status by immunohistochemistry involving relatively large sample size. Both methods similarly showed a large share of endocrine responsive tumors. Certain limitations have to be taken into account. Firstly, the cohort is a convenient hospital-based sample and does not reflect the true pattern within the population. Given the resource-limited setting, population-based sample collection is not feasible. Secondly, Ethiopian patients vary considerably linguistically, ethnically and culturally within the country. Therefore results cannot be generalized within the country and a nation-wide study investigating distribution intrinsic subtypes and further exploration of tumor biology is essential to capture the large number of different ethnic groups in Ethiopia which could possibly help to have a broaden understanding of BC biology.

留言 (0)

沒有登入
gif