The role of ultrasonographic findings for PIK3CA-mutated, hormone receptor-positive, human epidermal growth factor receptor-2-negative breast cancer

Abstract To determine whether ultrasound (US) features of breast cancer are associated with Breast Imaging and Reporting Data System molecular subtype, histologic grade, and hormone receptor status as well as to assess the predictive value of these features. Retrospective analysis of the medical records of 220 consecutive patients with invasive breast cancer was reviewed according to the PIK3CA-mutated molecular tumor subtype. US findings of all patients were analyzed. Breast tumors harboring a PIK3CA-mutation were large and exhibited liquefied necrosis and posterior echo attenuation in the nodule. Moreover, such tumors were lobulated and calcified. The aspect ratio of the PIK3CA-mutant was more likely >1. The average nodule elasticity (7.479 ± 0.993 m/s) was measured using US shear wave elastography. Microcalcification was easier to detect inside the nodule using a fluorescence technique. Measurement of the nodule blood flow spectrum showed that the internal blood flow resistance index of nodules was lower than that of other types of breast cancer. The sonographic features of PIK3CA-mutated breast cancers were strongly associated with extensive and liquefied necrosis. The ability to predict molecular subtypes, particularly using US to detect the triple-negative subtype, may play an important role in early management and treatment.


Introduction
Global cancer data released in 2020 by the World Health Organization International Cancer Research Institute included 2,261,419 new cases of breast cancer and 684,996 [1] deaths; including 416,471 new cases of breast cancer and 117,174 deaths in China (https://gco.iarc.fr/today/home).During the past 20 years, the development of experimental and clinically approved breast cancer-targeting drugs has been progressing rapidly.Examples include endocrine therapy, poly(ADPribose) polymerase (PARP) inhibitors (referred to as "PARPi"), phosphatidylinositol-3-kinase PI3K inhibitors, and human epidermal growth factor receptor-2 (HER2) inhibitors, which achieve good therapeutic effects [2][3][4][5][6].Moreover, despite intensive interest in immunotherapy, it does not effectively treat advanced breast cancer [6,7].Therefore, targeted therapy continues to play an important role in breast cancer chemotherapy.
Ultrasonography is the first choice for screening for breast cancer.The first aim of the ultrasound (US) examination is to identify benign and malignant nodules.Furthermore, with the rise of targeted therapy, it is critically important to identify the characteristics of PIK3CA mutations in US images of breast cancer.Therefore, our Radiology Department faces the most urgent key scientific problems such as analyses of the characteristics of US images of different genotypes of breast cancer and implementation of early diagnosis and treatment.This will be of great practical significance to expand the scope of clinical application of drugs and to lengthen patients' survival rates.
Therefore, the purpose of the present study was to determine whether the US features of breast cancer are associated with Breast Imaging and Reporting Data System (BI-RADS) molecular subtype, histologic grade, and HR status as well as their abilities to predict patients' outcomes.

Methods
The US features of 220 consecutive patients with primary invasive breast cancer, who were treated and followed at our breast cancer center between November 2011 and August 2013, were retrospectively evaluated using an electronic database.Of the 220 patients, we excluded 19 with incomplete data and bilateral and recurrent breast cancer.Thirty-three of 201 (16.4%) invasive breast cancers were multifocal, and in these cases, the largest lesion was evaluated.All patients had histologically proven breast cancer and molecular subtypes revealed by analyses of surgical specimens.Patients in the PIK3CA-mutant and non-mutant cohorts were randomized.
The Ethics Committee of the Beijing Chaoyang Hospital approved this study (2021-Ke-704).Informed consent from patients with PIK3CA-mutated breast cancer was not required by the Ethics Committee for this retrospective study.

Ultrasonography
US scans were performed using a 13-5 MHz linear transducer (VF13-5 Acuson Antares, Siemens, Erlangen, Germany), and the data were evaluated by two radiologists specializing in breast cancer with ≥5 years' experience performing breast imaging.Both radiologists were uninformed regarding the histopathology results.One radiologist assessed the US images of each tumor from the image archiving and communication system and soft copy images, and the second radiologist was consulted if a case was unclear.All US exams were performed by radiologists, and multiple images were recorded.
The database records of 201 consecutive patients with invasive breast cancer were reviewed according to the molecular subtype of PIK3CA-mutated disease.Sonographic tumor sizes were classified as ≤10 mm 3 and >10 mm 3 ; aspect ratios ≤1 and >1; internal echo as low, medium, and high; and ultrasonic shear wave elasticity (USWE) microcalcification as yes or none.Tumor margins were classified as circumscribed or non-circumscribed.The posterior acoustic features were divided into categories as follows: shadowing, enhancement, no change, and mixed pattern.Two senior attending physicians jointly evaluated the images.If there was disagreement, the chief physician was consulted.

Statistical analysis
SPSS 24.0 software (IBM, Chicago, IL, USA) was used for statistical analysis.Normally distributed clinical quantitative parameters are expressed as the mean ± standard deviation and compared using the Student's t test, while non-normally distributed variables are expressed as the median (interquartile range) and analyzed by using the Wilcoxon signed rank test.Data for categorical variables are expressed as numbers (percentages), and the chisquare test was used.Univariate and multivariate logistic regression models were used to analyze the relationship between the classifications.The receiver operating characteristic (ROC) curve was used to determine the optimal cutoff value of the C-statistic that best correlated with the patient's prognosis.The Kaplan-Meier method was used to estimate overall survival, and P < 0.05 indicates a statistically significant difference.

Results
Table 1 shows the tumor characteristics and US imaging features of patients included in the study.The mean age at diagnosis of patients with non-mutated tumors was 48 ± 15 years (range, 23-83 years), and the ultrasonographic mean tumor size was 22 ± 14.1 mm.Histologic grade was 1 or 2 in 73 patients (36.3%) and grade 3 in 128 patients (63.7%).Among 201 patients with invasive breast cancer, the molecular subtype was luminal A in 58 (28.9%), luminal B in 99 (49.3%),HER2-positive in 18 (9.0%),and triple-negative in 26 (12.9%).The research flow chart is shown in Figure 1.

US imaging features of PIK3CA-mutated tumors
The breast tumors of patients with mutated PIK3CA were large and exhibited liquefied necrosis.There was posterior echo attenuation in the nodule, and the tumors were lobulated and calcified.The aspect ratios of PIK3CA-mutated tumors were more likely ≥1.The average nodule elasticity was 7.479 ± 0.993 m/s, which was measured using USWE.

Discussion
In the present study, we determined whether the US features of breast cancer were associated with the BI-RADS molecular subtype, histologic grade, and HR status.Furthermore, we assessed the predictive value of these features.Our findings show that the sonographic features of PIK3CA-mutated breast tumors were obviously associated with large tumors and liquefied necrosis, suggesting the ability of US to predict molecular subtypes; particularly the triple-negative subtypes may play an important role in early management and treatment.Breast cancer is a heterogeneous disease with different histopathological and associated biological characteristics [1][2][3].Over the past few decades, treatment of breast cancer has entered the molecular, genomics, and proteomics levels, which has improved treatment efficacy to lengthen the survival of patient with breast cancer [4][5][6].Previously, breast cancer was treated as a disease without a classification.In 2000, the advent of molecular genetic technology made possible high-resolution molecular typing, culminating in personalized treatment [7][8][9].The 70-gene detection MammaPrint microarray was approved for clinical use in China, representing a very important milepost [1,8].The MammaPrint "70 gene detection" technique provides an important prognostic tool for ERPR-positive and HER2-negative patients with breast cancer, and the score is calculated at RNA transcriptional level [7][8][9][10][11][12][13][14].
Molecular biological studies reveal the heterogeneity of breast cancer, which help to obtain more precise prognostic and prediction information [10,11].Therefore, the molecular typing of breast cancer includes the phenotypes as follows: luminal A, luminal B, Her2 overexpression, and triple-negative breast cancer.Patients in the present study had luminal A and B molecular subtypes (78%).
Breast US has become an effective method to identify benign and malignant lesions, particularly those of young women with dense breast tissue [10][11][12].Although many studies focus on the ultrasonographic findings of breast  cancer grading, studies on the relationship between tumor subtypes are unavailable [3,8], and imaging data on the identification of molecular subtypes are unavailable as well.Compared with the HR-negative state, HR-positive tumors are more likely to have uncertain margins.In the present study, multivariate analysis revealed significant differences associated with age, grade, echoic, margins, aspect ratio, and microcalcification.Previously, most malignant breast tumors examined using US were thought to have posterior acoustic shadows with blurry edges [13,14].However, it is now well known that many tumors may exhibit variable posterior acoustic features.For example, recent studies show that a clear margin and post-enhancement are more likely to represent a higher level of tumor and receptor-negative status [13,15,16].Our results are consistent with these findings, particularly the statistically significant correlation between a clear margin and post-enhancement.
In the present study, histologic grade was 1 or 2 in 73 patients and grade 3 in 128 patients.Among 201 patients with invasive breast cancer, the molecular subtypes were luminal A in 58, luminal B in 99, HER2 in 18, and triplenegative in 26.Tumors with HR-positive status (n = 154) were more likely to have non-circumscribed margins (n = 108), and tumors with HR-negative status (n = 47) were more likely to have circumscribed margins (n = 32).Circumscribed margins also are among the most frequent alterations in solid tumors, identified in the cancers as follows: 42-55%, endometrial; 24%, cervical; 18%, colorectal, 13%, head and neck; and 12%, ovarian.Targeting the PI3K/mTOR pathway therefore may be particularly effective in cancers that signal through PI3Ka [17][18][19].Further studies with larger sample size are needed to investigate the differences in US features among patients with different expression levels of HER2 and HR.
The results of the present study showed that sonographic features of PIK3CA-mutated tumors were significantly associated with large and liquefied necrosis.Here we show posterior echo attenuation and microcalcification in the nodule and that the average nodule elasticity was 7.479 ± 0.993 m/s measured using USWE.Furthermore, measurement of the nodule blood flow spectrum reveals a lower internal blood flow resistance index of nodules  Ultrasonography for PIK3CA-mutated breast cancer  5 compared with those of other types of breast cancer.The ability use US to predict molecular subtypes, particularly the triple-negative subtype, may play an important role in optimizing early management and treatment.Large multicenter studies are required to confirm these findings.

Figure 2 :
Figure 2: US imaging features of PIK3CA-mutated breast tumors.(a) Large nodules with liquefaction necrosis.(b) Echo attenuation behind nodules.(c) Nodules lobulated with inner calcification.(d) Nodule aspect ratio >1.(e) Elasticity measurements of nodules through shear-wave techniques.(f) Intranodal microcalcifications detected using a fluorescence technique (microcalcifications in blue, white light).(g) Nodular blood-flow spectrum measurements show that the intranodular blood flow resistance index is lower than those of other types of breast cancer.

Table 1 :
Patients' tumor characteristics and US imaging features *Statistically significant.

Table 2 :
Logistic analysis results of features of PIK3CA-mutated breast tumors