A prediction model using 2-propanol and 2-butanone in urine distinguishes breast cancer

Shoko Kure,Sera Satoi,Nobuo Nakano,Masao Miyashita,Toshihiko Kitayama,Shinya Iida,Marina Yamada,Satoshi Miyashita,Noboru Uchiyama,Yuta Nagase,Hiroyuki Takei

doi:10.1038/s41598-021-99396-5

Abstract

Safe and noninvasive methods for breast cancer screening with improved accuracy are urgently needed. Volatile organic compounds (VOCs) in biological samples such as breath and blood have been investigated as noninvasive novel markers of cancer. We investigated volatile organic compounds in urine to assess their potential for the detection of breast cancer. One hundred and ten women with biopsy-proven breast cancer and 177 healthy volunteers were enrolled. The subjects were divided into two groups: a training set and an external validation set. Urine samples were collected and analyzed by gas chromatography and mass spectrometry. A predictive model was constructed by multivariate analysis, and the sensitivity and specificity of the model were confirmed using both a training set and an external set with reproducibility tests. The training set included 60 breast cancer patients (age 34–88 years, mean 60.3) and 60 healthy controls (age 34–81 years, mean 58.7). The external validation set included 50 breast cancer patients (age 35–85 years, mean 58.8) and 117 healthy controls (age 18–84 years, mean 51.2). One hundred and ninety-one compounds detected in at least 80% of the samples from the training set were used for further analysis. The predictive model that best-detected breast cancer at various clinical stages was constructed using a combination of two of the compounds, 2-propanol and 2-butanone. The sensitivity and specificity in the training set were 93.3% and 83.3%, respectively. Triplicated reproducibility tests were performed by randomly choosing ten samples from each group, and the results showed a matching rate of 100% for the breast cancer patient group and 90% for the healthy control group. Our prediction model using two VOCs is a useful complement to the current diagnostic tools. Further studies inclusive of benign tumors and non-breast malignancies are warranted.

Highlights

Abbreviations area under the curve (AUC) Area under the curve breast cancer patient (BCP) Breast cancer patient GCMS Gas chromatography–mass spectrometry healthy control (HC) Healthy control MG Mammography NA Not applicable ND Not determined Receiver Operating Characteristic (ROC) Receiver operating characteristic S.D
We assessed the potential of Volatile organic compounds (VOCs) to become biomarkers of breast cancer by constructing a prediction model using a training set and an external validation set with triplicated reproducibility tests
The BCPs and HCs were randomly allocated to a training set and an external set using software

Summary

Introduction

Abbreviations AUC Area under the curve BCP Breast cancer patient GCMS Gas chromatography–mass spectrometry HC Healthy control MG Mammography NA Not applicable ND Not determined ROC Receiver operating characteristic S.D. The early detection of breast cancer is an important step toward achieving efficient treatment. Mammography (MG), the most commonly used screening test at present, can detect breast cancers during the asymptomatic phase and reduce mortality among women of certain ages[2,3,4]. While the advantages of urine as an alternative matrix for volatile biomarkers have been outlined in lung c ancer[27], the data are scanty on cancers of other organs, including the breast. In this study we sought to identify and analyze VOCs that appear in the urine of breast cancer patients. We assessed the potential of VOCs to become biomarkers of breast cancer by constructing a prediction model using a training set and an external validation set with triplicated reproducibility tests

Methods

Results

Discussion

Conclusion