There really is no limit to the kinds of medical information that might be included to allow ever more precise feedback and guidance in the form of clinical-decision support.
—Clifford A. Hudis, MD
In June, Clifford A. Hudis, MD, Chief of the Breast Cancer Medicine Service and Attending Physician at Memorial Sloan-Kettering Cancer Center, and Professor in the Department of Medicine at Weill Cornell Medical College, will begin his term as President of ASCO. Among Dr. Hudis’ priorities will be to oversee the continued development and eventual deployment of ASCO’s breast cancer–specific prototype for CancerLinQ. (See “ASCO’s Approach to Health Information Technology and the Rapid-learning System,” in the January 15 issue of The ASCO Post.)
Recently, The ASCO Post talked with Dr. Hudis about CancerLinQ.
The CancerLinQ breast cancer prototype was unveiled at ASCO’s Quality Care Symposium this past November. What is the next step in the development of the prototype?
We have loaded de-identified data from breast cancer patients’ electronic medical records into the CancerLinQ system, so the next step is to begin to measure these clinical experiences in relation to ASCO’s (and other high-level) breast cancer clinical practice guidelines and ASCO’s Quality Oncology Practice Initiative (QOPI®) measures. [Four large cancer centers—Maine Center for Cancer Medicine, Marin Specialty Care and Marin General Hospital (California), Space Coast Cancer Center (Florida), and Tennessee Oncology—and IntrinsiQ LLC, a leading source of U.S.-based oncology data and analysis, are providing anonymous case data on approximately 100,000 patients for inclusion in the CancerLinQ breast cancer prototype.]
A critical point is that these patient records are being drawn from several widely used electronic medical records, emphasizing the ability of CancerLinQ to integrate medical information from many sources.
Is there a timeline for when CancerLinQ will be ready for clinical practice?
No, not yet, but we are making rapid progress. We are learning many lessons from the prototype. The next steps include accepting prototype data from additional sites. We are also building the data governance, privacy, and security standards for the full CancerLinQ system.
Will CancerLinQ have the ability to capture all the pertinent information about a patient, including molecular characteristics of the patient’s tumor, and then determine the most effective treatment options?
We hope the information the system will capture is even broader than that. The idea is to include all of the relevant information about a specific patient, which could include genomics, but it doesn’t have to in all situations. The CancerLinQ rapid learning system can also include any standard testing and testing yet to be developed that may be used in the future. It includes demographic information about patients, classic anatomic pathology reports, concurrent medications, and other illnesses the patient may have. Ultimately, the system should also include an assessment of the patient’s toxicity risk tolerance and goals of care.
There really is no limit to the kinds of medical information that might be included to allow ever more precise feedback and guidance in the form of clinical-decision support. But, yes, CancerLinQ will certainly include more and more genomic data primarily as a response to its demonstrated everyday utility, but eventually perhaps to contribute information that may aid in the exploration of important associations between specific genomic patterns and treatment and disease outcomes.
I think it helps to take a half-step back and recognize that these factors—and more—are the things that a physician would use to make some assessments about the best therapeutic approach for a patient. But in the end, this system is never going to treat a patient. It will look at a patient’s situation and determine what the standard therapy is to consider. If the physician elects to treat in a different fashion, the system will provide an alert detailing how he or she might be deviating from what is expected. But then it is still up to the doctor to be a doctor and, with the patient, determine the best course of care.
Similarly, the system might inform the doctor about a research opportunity that the doctor might want to consider for the patient. This has the exciting potential to both maintain high quality of care independent of geography and to increase the efficiency of clinical trials.
Will CancerLinQ have the potential to make the treatment decision-making process more precise?
It will include clinical decision-support technology that helps physicians identify the factors they need to make evidence-based treatment decisions and to also recognize factors that may have been omitted. A very simple example would be, if a patient has early-stage ER-positive breast cancer and there are no prescriptions for hormonal therapy entered into her electronic medical record, CancerLinQ would recognize that there was an omission and question why. Or more simply, it might say, “No prescription has been given for tamoxifen.” Then the physician might type in, “The patient has a specific contraindication and can’t take tamoxifen.”
It is important to emphasize that CancerLinQ is really just delivering the current standards in a faster, more efficient fashion. It is not ever replacing sound medical judgment and responsibility.
It also sounds like CancerLinQ has the potential to provide a greater opportunity for shared decision-making between the physician and patient.
That could be. Overall, the opportunity is for a better sharing and distribution of knowledge. Physicians don’t have to be at an academic or research center in order to be aware of the most recent treatment developments.
An example might be an oncologist seeing a patient who is 5½ years from diagnosis and has been on tamoxifen for 5 years. A few weeks ago the oncologist might have said, “You are done with treatment” and sent that patient for routine follow-up. If operational today, CancerLinQ could have provided information showing that a study presented at the most recent San Antonio Breast Cancer Symposium found that 10 years vs 5 years of tamoxifen has a survival advantage. Without having traveled to the meeting, this oncologist could be made aware of this finding to consider in the treatment of her patient today. There are myriad other examples, of course.
Do you know how expensive CancerLinQ will be for oncology practices and hospitals to adopt?
That’s a complicated question, because the actual computer network system is very expensive to build. But physician practices and cancer centers are already making investments in the creation of their own electronic medical records, so at the practice level there is not much more of an investment needed to implement this system.
CancerLinQ will be broadly compatible with existing electronic medical records, so it should not be a prohibitive expense at the practice level. In the end, it will have to offer our community an acceptable return on investment or it will not succeed.
Your term as ASCO President begins in June. What are some of your expectations, in terms of challenges and goals?
The one thing I’ve been consistently warned about by everybody is that whatever I think the challenges are, other things will happen that I don’t anticipate.
My goals, of course, include the continuing rapid deployment and maturation of CancerLinQ. For the longer term, one of the most critical issues for us is going to be the completion of an anonymized data set from the electronic medical records of the breast cancer patients for the CancerLinQ prototype and our ability to mine that data and use it productively.
That is why I think the implementation of CancerLinQ is so important. It isn’t just about having access to thousands of electronic medical records locked away in academic centers and oncology offices across the country. It is also having the ability to provide real-time, standardized, clinical decision-support tools built into the system, as well as other value-added tools, including QOPI performance measures. It is the potential of CancerLinQ to transform the way cancer is understood and treated that I think is so important about the technology. ■
Disclosure: Dr. Hudis reported no potential conflicts of interest.