Immune checkpoint inhibitors are a major breakthrough in the field of tumor treatment this century and have been approved in multiple tumor species. Currently, for immunotherapy, PD-L1 is a useful but imperfect biomarker. Whether more clues can be found at the genomic level is on

*For medical professionals reading only, reference

What are the specific mutation genes in different breast cancer subtypes? What is the relationship between gene mutations and ICI biomarkers? How is PD-L1 related to Ki67 or TMB?

immune checkpoint inhibitor (ICI) has been approved in multiple tumor species as a major breakthrough in the field of tumor treatment this century. Recently, the my country Drug Administration also approved the indications of immunotherapy for for neoadjuvant therapy for triple-negative breast cancer, officially opening a new era of breast cancer immunotherapy in my country [1]. However, while immunotherapy brings benefits, it also has the problem of a low proportion of benefiting people and an increase in toxic and side effects. Therefore, how to screen people who truly benefit is a key issue that needs to be solved urgently. Currently, for immunotherapy, PD-L1 is a useful but imperfect biomarker . Whether more clues can be found at the genomic level is one of the current research directions [2]. In breast cancer, previous research on immunotherapy has mainly focused on triple-negative breast cancer. It is also worth paying attention to how other subtypes perform in PD-L1 and related markers [3].

Guangdong Provincial People's Hospital Professor Liao Ning led a team to conduct a study that explored the genetic altered characteristics and PD-L1 levels of different breast cancer subtypes in the Chinese population and evaluated their correlations. The study not only reveals the association between different biomarkers of immunotherapy, but also provides more possibilities for the treatment of different subtypes of breast cancer. Recently, the research was officially published in the authoritative journal of Cancer Medicine [4]. The key content is sorted out as follows for readers.

immune checkpoint inhibitor (ICI) is becoming a novel therapeutic tool for advanced breast cancer, and the discovery of selective ICI targeting programmed cell death ligand 1 (PD-L1) has created new opportunities for breast cancer treatment. Previous studies have shown that the expression of PD-L1 in different breast cancer subtypes has diversity , and different breast cancer subtypes respond differently to PD-L1 inhibitors. However, PD-L1 has limitations as a predictive biomarker for breast cancer, and exploring the correlation between PD-L1 expression levels and other biomarkers in breast cancer may become a promising strategy. Furthermore, given that certain gene mutations may affect the ability of tumor cells to evade immune surveillance, it is necessary to explore whether genomic changes in breast cancer are associated with existing ICI biomarkers (such as PD-L1).

study results

patient characteristics

From December 2017 to July 2019, a total of 301 patients with primary breast cancer in Guangdong Provincial People's Hospital were admitted, including 16 ductal carcinoma in situ (DCIS), 251 invasive ductal carcinoma (IDC) and 34 other histological types. The clinical characteristics of 301 patients with breast cancer are shown in Table 1. The median age of the patient was 48 years (range: 23-83 years). The population consisted of stage I, II, III and IV patients, accounting for 22.92%, 41.86%, 13.29% and 8.64% respectively. There were 137 (45.51%) patients with breast cancer who were PD-L1 positive, of which 32.89% (1≤CPS 10) were defined as PD-L1-L, 12.62% (CPS ≥10) were defined as PD-L1-H, and Ki67-H was found in 75.08% of cases. According to the immunohistochemical staining results, the breast cancer patients in the study cohort were mainly divided into HR+/HER2-, HR+/HER2+, HR-/HER2+ and HR-/HER2- subtypes, accounting for 61.46%, 15.62%, 12.62% and 10.30%, respectively.

Table 1. Cohort baseline characteristics (N = 301)

Genetic alteration characteristics of Chinese breast cancer patients

3892 clinically related variants were found in these 301 breast cancer samples in the Chinese population. TMB-H was observed in 7.64% of cases, compared with 92.36% of TMB-L. Of all changes, 1781 (45.76%) was replacement/insertion, 1426 (36.64%) was gene amplification, 404 (10.38%) was truncated, 252 (6.47%) was fusion/rearrangement, and 29 (0.75%) were gene homozygous deletion. The most common variants found in the Chinese breast cancer cohort were TP53 mutations (54.5%), followed by PIK3CA (39.5%), ERBB2 (30.2%), CDK12 (22.9%), and GATA3 (18.3%) mutations. Other top mutant genes are shown in Figure 1.

Figure 1. Mutation characteristics of breast cancer, genomic alteration spectrum of breast cancer

The mutation characteristics of different breast cancer subtypes are different, see Supplementary Figure 1.The most common mutant genes in the HR+/HER2-subtype were PIK3CA (40.5%), TP53 (33.5%), GATA3 (24.9%), CCND1 (17.3%), and FGF19 (15.7%) (Figure 2A). The most common mutant genes in the HR+/HER2+ subtype were ERBB2 (89.4%), TP53 (74.5%), CDK12 (66%), PIK3CA (34%), and RARA (31.9%) (Figure 2B). The most common mutant genes in the HR-/HER2+ subtype were TP53 (97.4%), ERBB2 (94.7%), CDK12 (73.7%), PIK3CA (55.3%) and NF1 (31.6%) (Figure 2C), and the most common mutant genes in the TNBC subtype were TP53 (96.8%), PIK3CA (22.6%), PTK2 (22.6%), KMT2C (19.4%) and FAT4 (16.1%) (Figure 2D). TP53 (p = 4.81E-23), ERBB2 (p = 1.57E-48), and CDK12 (p = 2.67E-30), TOP2A (p = 1.65E-05), SPOP (p = 0.00016), RARA (p = 3.74E-06), PIK3CA (p = 0.0402), PTK2 (p = 0.00082), and FAT4 (p = 0.0400) were statistically significant among the four breast cancer subtypes (Fig. 3), and other genes in these four subtypes are in Supplementary Table 1. Moreover, several genes were found to be mutated in a single subtype (Supplementary Table 1). DDR2 (5.26%) and MYCL (5.26%) were mutated only in the HR-/HER2+ subtype, and H3-3A (9.68%) and NRAS (6.45%) were associated with the HR-/HER2- subtype. Differential mutant genes in different subtypes may provide potential guiding significance for future treatments.

Figure 2. Mutational changes observed in different breast cancer subtypes. (A) The 10 most frequent mutations in the HR+/HER2-, (B) HR+/HER2+, (C) HR-/HER2+ and (D) TNBC subtypes change

Figure 3. Mutation genes with statistical differences between breast cancer subtypes. (A-I) Mutation frequencies of TP53, ERBB2, CDK12, TOP2A, SPOP, RARA, PIK3CA, PTK2 and FAT4 were shown in the breast cancer subtype. P 0.05 is considered to be statistically significant

Supplementary Figure 1. Mutation characteristics of different subtypes. (A) Genomic alteration characteristics of (A) HR-/HER2+ subtype, (B) HR-/HER2- subtype, (C) HR+/HER2+ subtype and (D) HR+/HER2- subtype

Supplementary Table 1. Relevant PD-L1 expression of different mutation genes of 4 breast cancer subtypes

PD-L1 expression and clinical factors and genomic changes

Among the enrolled patients, 164 (54.5%) breast cancer were PD-L1 negative (CPS1), while 137 (45.5%) breast cancer was PD-L1 positive (CPS≥1), including 99 (72.3%) PD-L1-L and 38 PD-L1-H (27.7%) (Table 1). The distribution of PD-L1-negative, PD-L1-H and PD-L1-L patients in different breast cancer subtypes was further analyzed. The percentage of patients with positive PD-L1 expression was higher in the HR-/HER2-subgroup (70.9% vs 38.4%, p = 0.0037), mainly due to the percentage of PD-L1-H levels (35.5% vs 9.2%, p = 0.0004). The PD-L1 positivity rate was the lowest in the HR+/HER2- subgroup (38.4%), while the PD-L1-H percentage was the lowest in the HR+/HER2+ patients (4.26%) (Fig. 4C). In addition, PD-L1 expression in the entire HR+, HR-, HER2+ and HER2- subgroups were compared (Fig. 4D). A significantly higher PD-L1-H frequency was observed in HR-patients compared with HR+ patients (27.5% vs 8.2%, p = 0.0001), while no difference was found between the HER2- and HER2+ groups. The

study also identified the relationship between PD-L1 and clinicopathological characteristics and observed that PD-L1 levels were significantly correlated with ER, PR status, and histological grading (Table 2). The mutation characteristics of PD-L1-positive (CPS ≥ 1) and PD-L1-negative breast cancer are shown in Figure 4A. The most common mutation types were substitution/insertion (45.49%), followed by gene amplification (36.69%), truncation (11.0%), fusion/rearrangement (6.13%), and gene homozygous deletion (0.69%; Figure 4B). Compared with PD-L1-L and PD-L1-negative breast cancer, there were more substitutions/insertions in PD-L1-H breast cancers (p = 1.55E-07), but less gene amplification (p = 2.48E-08) (Fig. 4E). This difference was also observed in the mutation types and frequency of subpopulations of PD-L1-H, PD-L1-L and PD-L1-negative breast cancer. The top 5 genes in the PD-L1-H subset were TP53, PIK3CA, ERBB2, CDK12 and NF1, followed by TP53, PIK3CA, ERBB2, CDK12, and GATA3 (Supplementary Fig. 2). In addition, TP53 (84.21% vs. 61.62% vs. 43.29%, p=3.778E-06), MYC (13.16% vs. 3.03% vs. 3.66%, p=0.047), FAT4 (13.16% vs. 8.08% vs. 1.22%, p=0.0009), PBRM1 (7.89% vs. 0.0% vs. 0.61%, p=0.007), PREX2 (7.89% vs. 1.01% vs. 1.22%, p=0.007), The highest mutation frequency of 0.035) was significantly different in the PD-L1-H cohort and the PD-L1-L and PD-L1-negative cohorts (Fig. 4F).

Table 2. Comparison of clinical and molecular characteristics of patients with negative and positive PD-L1 expression

CPS 1 was defined as Low; 1 ≤ CPS 10 was defined as Middle; CPS ≥ 10 was defined as High.

Figure 4. Heat map of mutant genes compared according to different PD-L1 levels. (A) Genome altered profiles were shown based on CPS. (B, E) Special changes were observed among different breast cancer subtypes. (C) Distribution of PD-L1-negative, PD-L1-H and PD-L1-L in each subtype of breast cancer. (D) Distribution of PD-L1 negative, PD-L1-H and PD-L1-L in HR-, HR+, HER2- and HER2+ cohorts. (F) Statistical differences in between groups based on PD-L1 expression levels. P 0.05 is considered to be statistically significant

Supplementary Figure 2. The top 10 common mutant genes in PD-L1-negative and PD-L1-positive were identified. (A) The top 10 common mutant genes (CPS ≥10) in the PD-L1 high expression cohort. (B) The top 10 common mutant genes in the PD-L1 low expression cohort. (C) The top 10 common mutant genes in the PD-L1 negative cohort (CPS 1)

The correlation between PD-L1 and Ki67 and TMB

TMB has become a predictor of guiding immunotherapy in real clinical practice, and has attracted great attention from scholars in the field. First, this study found that the percentage of TMB-H in the HR+/HER2- and HR-/HER2+ subgroups (9.2% and 10.5%, respectively) was slightly higher than that of the HR-/HER2- and HR+/HER2+ subgroups (3.2% and 2.1%, respectively), but no statistical difference (Fig. 5A). The researchers then analyzed the relationship between TMB and clinical factors and proved that the status of postmenopausal was significantly correlated with TMB-H (p = 0.018, Table 3). The study also found positive correlation between TMB and PD-L1 in the HR+/HER2- cohort (p = 0.0074), HR+ cohort (p = 0.0046), HER2- cohort (p = 0.023), and all breast cancers (p = 0.007) cohorts (Figure 5B-D, Supplementary Fig. 3A, and Supplementary Table 2), indicating that the percentage of TMB-H breast cancer in PD-L1-positive breast cancer was higher than that in PD-L1-negative breast cancer. In the HER2-cohort, there were more patients with PD-L1-H than patients with PD-L1-L (p = 0.032).

Table 3. Comparison of clinical and molecular characteristics of TMB-high and TMB-low patients

Table 2. Correlation analysis of PD-L1 and TMB

Figure 5. Correlation between TMB and Ki67 and PD-L1 in different subtypes. The percentage of TMB-H patients was determined in (A) HR+/HER2-subtype and (B) all breast cancer patients with different CPS. The percentage of Ki67-H patients was determined in (C) HR+ subtype, (D) HER2- subtype, and (E) breast cancer patients with all different CPS. It is believed that p 0.05 has statistically significant

Ki67 index (about ≥20%) is related to the malignant phenotype and adverse prognosis of breast cancer. The study demonstrated that the high Ki67 index was negatively correlated with ER(p = 1.25E-26) and PR states (p = 3.36E-07) and positively correlated with the HER2 state (p = 0.0008) and histological grading (p = 0.0005, Table 4). Furthermore, the proportion of Ki67-H patients in PD-L1-H patients was higher in the HR+/HER2- cohort (p = 0.001), HR+/HER2- cohort (p = 0.002), HR+ cohort (p = 0.004), HER2+ cohort (p = 0.001), HER2- cohort (p = 0.0003), and all breast cancer patients (p = 0.0006), but not related to TMB levels (Fig. 5E-I, Supplementary Fig. 3B, Supplementary Table 3, and Supplementary Table 4). These results may reveal more possibilities for the treatment of different subtypes of breast cancer.

Table 4. Comparison of clinical and molecular characteristics of Ki67-H and Ki67-L patients with

Supplementary Figure 3. Relationships of TMB, Ki67 and PD-L1 in different subtypes. (A) Percentage of TMB-H patients were measured in all breast cancer patients with different PD-L1 levels. (B) Percentage of Ki67-H patients were determined in all breast cancer patients with different PD-L1 levels. It is believed that p 0.05 is statistically significant

study discussion

This study systematically analyzed the correlation between genome and PD-L1 to determine the relationship between gene mutations and ICI biomarkers. The study explored mutational changes based on PD-L1 levels and identified specific mutant genes in different breast cancer subtypes. In addition, the study observed the correlation between PD-L1 and Ki67 or TMB, which helped introduce more ICI links with other factors.

Breast cancer is a heterogeneous disease with different clinical outcomes and treatment responses for different subtypes. Different breast cancer subtypes can be identified by different biomarkers with different clinical characteristics that will affect the treatment of patients.The most common PIK3CA, MAP3K1 and TP53 mutations were observed in the cancer genome map network; the most common TP53 and PIK3CA mutations in the HR+/HER2-subtypes were observed; the most common TP53 and PIK3CA mutations in the HER2+subtypes were observed, and the most common TP53 mutations in the basal-like type were observed. In the Chinese cohort, the most common PIK3CA, TP53 and GATA3 mutations were found in the HR+/HER2-subtype; the most common ERBB2, TP53 and CDK12 mutations were found in the HR+/HER2+ subtype; the most common TP53, ERBB2 and CDK12 mutations were found in the HR-/HER2+ subtype; the most common TP53, PIK3CA and PTK2 mutations were found in the TNBC subtype. Most notably, the mutation frequency of TP53, ERBB2, CDK12, TOP2A, SPOP, RARA, PIK3CA, PTK2, and FAT4 was statistically significant among the four breast cancer subtypes. In the near future, molecular testing will help to better understand disease biology and tailor treatment strategies for each specific individual.

immunotherapy is one of the most encouraging cancer treatment findings in recent years. Breast cancer characteristics that may be associated with ICI response include higher mutation load, higher PD-L1 levels, and enhanced TIL. Immunomodulation/immunotherapy (PD1 or PD-L1 antibody) is a good choice for patients with PD-L1 positive status, and PD-L1 has been reported as an active immune checkpoint for a variety of cancers, including breast cancer. This study compared genomic changes in PD-L1-positive and PD-L1-negative patients. The mutation frequency of TP53, MYC, FAT4, PBRM1, and PREX2 was observed in the PD-L1-H cohort of breast cancer than that of the PD-L1-L and PD-L1 negative cohorts. PD-L1-positive patients with these mutant genes may be more efficient in immunotherapy or obtaining a better prognosis.

PD-L1 has some limitations as a biomarker because of its dynamic and heterogeneous expression in the tumor microenvironment, variable detection and interpretation, and lack of standardization between platforms and tumor types. The TMB observed in metastatic breast cancer is higher than that in primary tumors and is also shown to be associated with HR-breast cancer and HER2+ breast cancer. However, the predictive value of TMB for breast cancer immunotherapy remains controversial. Several studies have observed that PD-L1 and TMB are independent predictors of immune checkpoint blockade (ICB) response, respectively, with low correlation between PD-L1 levels and TMB in a variety of tumor types. The landmark IMpassion130 trial showed that TMB predicts better benefits for ICB in PD-L1-positive patients. Given the limitations of PD-L1 as a biomarker, it is very meaningful to determine the role of TMB in predicting anti-PD-1/PD-L1 therapeutic responses. In addition, this study observed a positive correlation between TMB and PD-L1 in the HR+/HER2-subtype, and showed a higher proportion of Ki67-H patients in the HR+/HER2- and HR+/HER2+ cohorts, but not related to TMB levels. The relationship between PD-L1, TMB and Ki67 predicted values ​​may provide more possibilities for the treatment of different subtypes of breast cancer. Importantly, more clinical trials need to be designed to examine clinical results based on PD-L1 levels, Ki67 and/or TMB combinations, which, when used in combination, can identify different populations that cannot be identified when used alone.

study conclusions

This study analyzed mutation changes in patients with HR+, TNBC and HER2+ subtype breast cancer, and determined special changes between different subtypes. In addition, the study explored mutational alterations based on PD-L1 levels and identified specific mutant genes in different breast cancer subtypes. Moreover, the correlation between PD-L1 and TMB and Ki67 was also analyzed, which helps introduce new treatments to treat specific patients with subgroup . Molecular testing will help better understand disease biology and tailor treatment options for specific individuals. Verification exploration of the use of comprehensive predictive biomarker selection checkpoint inhibitors as combination therapy options for breast cancer is underway.

Expert profile

Liao Ning Professor

MD, professor, doctoral supervisor

Administrative director of the Department of Breast Surgery in Guangdong Provincial People's Hospital

Director of the International Council of American Association of Oncology Surgeons (SSO)

Director of the International Sentinel Lymph Node Association (ISNS) International Council

Member of the Expert Group of the United States NCCN Breast Cancer Guidelines (Chinese Version)

Member of the Expert Group of the National Health Commission Medical Administration Department "Breast Cancer Treatment Specifications" Member

National Health Commission's "Guidelines for Breast Cancer Diagnosis" Expert Group Member

National Health Commission's Expert Committee on Rational Drug Use of National Health Commission's "Tumor Drug Group" Expert Group Member

References

[1].

https://www.msdchina.com.cn/company_news_2022-11-09

[2].Fountzilas E, Kurzrock R, Vo HH, Tsimberidou AM. Wedding of Molecular Alterations and Immune Checkpoint Blockade: Genomics as a Matchmaker. J Natl Cancer Inst. 2021;113(12):1634-1647.

[3].Agostinetto E, Gligorov J, Piccart M. Systemic therapy for early-stage breast cancer: learning from the past to build the future. Nat Rev Clin Oncol. 2022;19(12):763-774.

[4].Li K, Cao L, Li C, et al. Genomic alteration profile and PD-L1 expression among different breast cancer subtypes in Chinese population and their correlations.

Cancer Med. 2022;10.1002/cam4.5314.

This material is provided by AstraZeneca and is only for medical and health professionals to communicate in medical science, and is not used for promotional purposes.

Approval number: CN-106676

Validity period: 2023-11-28

*This article is only used to provide scientific information to medical personnel, and does not represent the views of this platform.