Cancer... the dreaded C word strikes fear in the hearts of all. Cancer is the second leading cause of death globally. In India, the most recurrent forms of cancer to afflict the population include lung, breast, cervical and colorectal. Of these, breast cancer is among the most prevalent in women with 1.6 lakh new cases every year accounting for 14% of all cancers in India. The mortality is also quite high with around 87000 women dying of breast cancer in India every year. These may be mere statistics for most of us, but these numbers were life changing for Dr. Manjiri Bakre.

During her stint as a doctorate student at the Indian Institute of Science, Dr Bakre's close friend was diagnosed with early stage breast cancer (usually Stage I or II). She underwent surgery and thought that she was on the road to recovery. Unfortunately, the cancer recurred and metastasised (spread to other organs in the body). Despite aggressive treatment, Dr Bakre's friend succumbed to the illness. This incident led Dr Bakre to wonder if there could be better way to understand cancer and the way it takes over a host's body. Could there be a way to gauge a patient's course of treatment based on the aggressive biology of the tumor?

"Companies in the West were developing tests that could analyze tumour biology and determine the risk of relapse in patients with early-stage breast cancer patients. These tests could help personalize a patient’s treatment plan based on that patient’s individual risk of relapse. Since these tests are very expensive, I approached doctors in India and pitched my idea of developing an indigenous and affordable version of such a test," says Dr. Bakre. This led to the conceptualisation of OncoStem in 2011. The focus of the company was building tests to predict cancer relapse, to help personalized cancer treatment planning, inspired by cancer stem cells, which are known to play a role in cancer relapse.

CanAssist Breast: The Flagship Product

OncoStem flagship product “CanAssist Breast” is a machine learning-based prognostic test that helps personalize treatment for early stage breast cancer patients who are hormone receptor positive (HR+) and HER2-negative. The standard treatment for HR+ HER2-negative breast cancer involves both chemotherapy and hormone therapy. The risk of relapse in early stage (Stage I and II) HR+ HER2-negative breast cancer is low (~15-20%) even if patients are given hormone therapy alone. This implies a majority of patients (nearly 80%) are being overtreated with chemotherapy, which has toxic side effects and lowers quality of life of patients. 

"We use a cutting-edge machine learning-based algorithm which predicts the risk of recurrence for every patient. Machine learning is known to be a more advanced tool to develop prognostic tests where patterns of patient information need to be understood and analysed. It maximizes diagnostic accuracy thereby improving patient outcomes. Machine learning based methods also have flexible “transfer functions” which allow them to model complex processes such as tumour recurrence. Other competitor tests use older statistical techniques which are not designed to maximize accuracy and therefore are an outdated choice for complex diseases like cancer," she says.

CanAssist Breast is an affordable option that can help identify which breast cancer patients can be spared of chemotherapy. OncoStem has worked with about 12 hospitals in India, to develop and validate CanAssist Breast. Tumor tissues were analysed for multiple biomarkers/proteins reflective of aggressive biology of the tumor. Using this data support vector machine (SVM) based statistical model was developed which assigns a ‘risk score’ based on the key selected five biomarkers and clinicopathological information for each patient. The ‘risk score’ is indicative of risk of recurrence. The ‘risk score’ categorizes patients based on the ‘risk of cancer recurrence’ clearly as either ‘low or high’. Thus, by analysing patient’s tumor CanAssist Breast ‘identifies’ patients who will have minimal benefit (Low-risk) or who will benefit the most (high risk) by adding chemotherapy to their treatment. Patients classified as low-risk can potentially avoid chemotherapy, its costs and side-effects. This algorithm was validated on clinical samples from Indian, US and European patients in a multi-centric validation study. Various quality related and regulatory approvals were applied for given that this is a medical test, before the test was commercially launched in 2016. CanAssist Breast is performed in OncoStem’s NABL (ISO15189) and CAP accredited laboratory in Bangalore. The product is also ISO 13485 certified and CE marked. The combination of biomarkers analyzed is patented, the US and India patents have already been granted, and patents are pending in various other geographies

"CanAssist Breast so far has helped 70% of early-stage breast cancer patients avoid chemotherapy. Not only did these patients avoid the toxic side effects but also the financial toxicity of chemotherapy. 93% of doctors complied the test recommendation of avoiding chemotherapy, showing us that physician confidence in the test is high," adds Dr. Bakre. Currently, OncoStem is working on a solution on another subtype of breast cancer as well as ovarian cancer. The team is also working on full automation of the test. They claim to use an auto-stainer for the immunohistochemistry. "We are working on digital pathology solutions that will allow complete automation and also decentralization. This will increase throughput and also allow any hospital in the world to conduct the testing in their own laboratory."

When Healthcare Data Meets Machine Learning: Challenges, Workarounds & Reliability Issues:

Unlike developed countries, conducting validation studies in India can be challenging since clinical trial frameworks are not as well established. Dr Bakre approached nearly 60 hospitals to work on developing and validating CanAssist Breast, which ultimately led to ~12 hospitals signing up. She believes there are many reasons for this - one is that the doctor-patient ratio in India is well below the developed world. Each oncologist in India sees many more patients per day leaving them with little time to focus on research studies. Secondly, lack of documented patient records and follow-ups also makes it difficult to conduct studies that require patient history. Patients often move back to their hometowns after treatment and are lost to follow-up. Availability of data is crucial to AI/ML. A machine learning algorithm can only be as good as the input data. Robust data makes for a robust algorithm, otherwise its garbage in, garbage out. And finally, there are factors like lack of enforcement and encouragement in research studies, clinical trials, detailed documentation from government and regulatory authorities.

"Specifically, in healthcare, machine learning has led to exciting new developments that could redefine cancer diagnosis and treatment in the years to come. ML can increase access to treatment in developing countries which don’t have enough cancer specialist doctors to deal with the rising incidence of cancer that can treat certain diseases, it can improve the sensitivity of detection, add more value in treatment decisions of cancer, and it can help personalize treatment so that each patient gets the treatment that’s best for them. In many cases they can even add to workflow efficiency in hospitals. The possibilities are endless," says Dr. Bakre.

She explains further with some examples - LYNA (LYmph Node Assistant) by Google detects spread of breast cancer metastasis early and can reduce the burden on Pathologists as well. A deep learning convolutional neural network or CNN - developed by a team from Germany, France and the USA can diagnose skin cancer more accurately than dermatologists. In a recently reported study, the software was able to accurately detect cancer in 95% of images of cancerous moles and benign spots, whereas a team of 58 dermatologists was accurate 87% of the time. The move from lab to actual practice has happened already for some AI-based solutions such as the FDA-approved imaging tool called IDx-DR for diagnosing diabetic eye disease.

"Technologies like AI indeed have tremendous potential and all stakeholders like the promising algorithms, accurate clinical and relevant in vivo data, clinicians, institutions have to align themselves to reap meaningful benefits from it. One must remember that excellent technical innovations in AI can not fix social/political problems. Also, the data input to AI must be in high volume and of clinically high quality/relevance. Fundamentally flawed data cannot substitute for high volume. Currently most of the AI applications are using the paradigm of ‘deductive reasoning’ and we need to move from there towards ‘inductive reasoning’. We have travelled fair amount in the AI path to excellency, but one must be cautious going further to embrace the brilliant promise it holds. What we need next is to move from theoretic benefit and evangelical sales to established use cases and robust, clinically-relevant data."

Made in India and Made for India

CanAssist is a classic example of an indigenous product for Indians. It is the only test to be validated on Indian patients. Competitors have developed products and got them validated in Western countries, mostly on Caucasian patients. Caucasian breast cancer patients tend to be older and post-menopausal at diagnosis, whereas their counterparts in India tend to be younger and pre-menopausal. The USP of CanAssist is that this critical data point has been taken into consideration from the initial stages of test development. "Our offering is the only product to have data on Indian patients. We are also the most affordable offering in the market today, giving us another edge over prohibitively expensive Western tests, " adds Dr. Bakre. Moreover, the turnaround time is the shortest - the results are reported in 8-10 days compared to nearly 3 weeks, which is market standard.

Future Plans:

In addition to automation of CanAssist Breast, OncoStem is working on expanding the access of their flagship product to South East Asia and Middle East. They are also appending the test menu for prognosing two more cancers and partnering with hospitals across India, US and EU for clinical validation of these tests.