A measured approach

How long should it take to get a breast cancer diagnosis?

According to a working group of the Canadian Breast Cancer Screening Initiative, when an abnormality is seen on a mammogram, 90% of women should get answers within five weeks. That’s if a biopsy, where cells from a lump are examined under a microscope, isn’t needed. If a biopsy is needed, then women should get resolution within seven weeks.

How are we doing? A 2014 report by the Canadian Partnership Against Cancer that looked at wait times for breast cancer diagnosis in Canada states that Ontario, along with the rest of the provinces, is not meeting this target. Only the Northwest Territories meet the wait time target in cases where a tissue biopsy is required.

Researchers at Sunnybrook Research Institute (SRI) are developing an imaging tool that could one day give women and their health care providers the answers they need much sooner. Leading the team are Dr. Ali Sadeghi-Naini, a scientist at SRI, and Dr. Greg Czarnota, director of the Odette Cancer Research Program at SRI. They published a paper in Scientific Reports that showed this tool, based on quantitative ultrasound, could accurately distinguish between benign and cancerous lesions.

“One of the benefits of this technique is that it’s almost real-time and it’s accessible. Patients usually wait weeks for a biopsy or MRI results, but with this technique we could very quickly triage critical cases,” says lead author Sadeghi-Naini.

In it, 78 women with suspicious breast lesions were recruited through the Rapid Diagnosis Unit at Sunnybrook’s Odette Cancer Centre. The researchers used quantitative ultrasound to classify benign lesions and malignant tumours. They compared results of the technique with analyses of biopsy samples by pathologists and studies of MRI scans by radiologists. Using a hybrid quantitative ultrasound biomarker, they could distinguish between benign and cancerous lesions with more than 90% accuracy.

How does the technique work? The researchers used the raw data generated by ultrasound machines to analyze features of the ultrasound signal. The abnormal make-up of cancerous tissue tends to result in a higher-intensity signal, notes Sadeghi-Naini. “Tumour cells grow much faster and in an irregular manner. Usually the cells are dense and very tightly spaced, and the vasculature is very irregular. Potentially, the power of the echo we get from ultrasound is stronger.”

Based on measurements of the return of the ultrasound signal throughout the lesion, the researchers are able to model or “map” its microstructure, which includes cells and blood vessels. Furthermore, by a method called texture analysis, they can show where in the lesion there are differences in signal characteristics that are based, for example, on changes in the size, density and spacing of suspicious cells, which give the images “texture.”

In a previous study of women with locally advanced breast cancer, Sadeghi-Naini and colleagues showed that quantitative ultrasound accurately shows a patient’s response to chemotherapy as early as one week into treatment. They are excited at the possibility of using the technique to screen for breast cancer. Dr. Belinda Curpen, a radiologist in Sunnybrook’s Breast Imaging Centre who works in the Rapid Diagnostic Unit (RDU) and a co-author of the study, highlights its potential clinical utility. She notes that for women who come to Sunnybrook with a breast lump, the testing protocol is as follows: mammography, ultrasound and, if necessary, a biopsy. It takes a few weeks for an appointment, and another week to complete imaging tests and analysis if the patient isn’t in the RDU. When all is said and done, it can take about a month to find out whether a lump is cancerous. Quantitative ultrasound-based imaging could improve efficiency. “Let’s say we could use this in our day-to-day practice, we would be able to tell without doing a biopsy if something is malignant or not. We could reduce a lot of the biopsies that we’re doing for lesions that are obviously benign on quantitative ultrasound [imaging].”

Within the RDU, doing away with needless biopsies would enable more patients to be seen. On average, it cares for nine patients per week, notes Curpen. The limitation in providing next-day diagnosis is partly related to the availability of radiologists and also the fairly small number of biopsies that can be processed quickly by the pathology department.

Moreover, knowing right away if a tumour is malignant would expedite treatment. For instance, outside the RDU, it takes a week to get a pathology report based on biopsy results. If that report confirms someone has breast cancer, then there is another test to determine whether the tumour depends on hormones like estrogen and progesterone to grow. Chemotherapeutic drugs are prescribed accordingly. If, however, a tumour were to be classified as cancerous using quantitative ultrasound, then hormone receptor testing could immediately follow. “If we could do that right at the patient’s appointment and say, ‘OK, this is cancer, and this is likely to respond to this type of chemo,’ that would be amazing,” says Curpen, who cautions that further validation of the technique is required.

Looking ahead, Sadeghi-Naini says the long-term goal is to replace unnecessary pathological biopsies that use a needle with acoustic biopsies using quantitative ultrasound. The medical implications, says William Tran, a radiation therapist at Sunnybrook and a co-author of the study, are that “you could provide a more rapid diagnosis through which you could then pathway the patients to the most appropriate care in a timely manner. That’s really what we’re hunting down here.”

This research was funded by the Breast Cancer Society of Canada, Canadian Breast Cancer Foundation, Canadian Institutes of Health Research, Natural Sciences and Engineering Research Council of Canada, and Terry Fox Foundation.

Original article: Sadeghi-Naini A, Harini S, Tran WT, Hadizad F, Bruni G, Rastegar RF, Curpen B, Czarnota GJ. Breast lesion characterization using textural features of quantitative ultrasound parametric maps. Sci Rep. 2017 Oct 20;7(1):13638. doi: 10.1038/s41598-017-13977-x.