PrecisionProfile develops platform to view cancer data
Service normalizes data from several sources, enabling oncologists to focus on treatment, says Dave Parkhill.
As oncologists begin to draw more data sources into their efforts to more precisely treat the diseases in patients, their research efforts are becoming more complex. A Boulder, Colo.-based company is working to build a platform to aid their efforts.
PrecisionProfile is designing a data normalization effort that intends to take those various types of data and speed the process of treating patients. The company expects its efforts will remove some of the drudgery that clinicians now face, facilitate they research they need to do and speed treatment for cancer patients.
Effectively treating these patients requires oncologists to know more than simply in which organ the cancer has its origins. Now, optimizing treatment requires them to know the specific genomic information that precipitated the disease.
“Clinicians now are at the point that the question is now what is the underlying genomic mutation describing the cancer,” says Dave Parkhill, CEO of PrecisionProfile. Then, knowing that, the clinicians may want to use that knowledge to determine what is currently known to be the most effective treatment for patients with that genetic profile for that specific variant of cancer.
Also See: How precision oncology will use data to advance cancer treatment
Delivering those insights is difficult because a patient’s datasets can be scattered in a variety of locations and in several different types of formats. Those researching treatments “are instead wasting precious hours preparing and massaging the data before analysis can even be conducted,” Parkhill says.
For example, cancer genomic data—comprised of a molecular profile of patient cells and chromosomes—is in VCF files and includes a long list of tags and genomic variants for each patient, and the file size can be enormous, with one person’s DNA taking up as much as 50 gigabytes. Demographic and clinical data is typically in XML format, while genetics data can be in diverse formats, often varying depending on the type of cancer.
PrecisionProfile aims to speed the treatment research process by enabling researchers to upload patient-specific information, normalize variances in information, then be able to match patients with similar genetic characteristics and cancers to find specific drugs and treatment protocols that are the most effective.
The company is using a product from Paxata to support the analytics platform. Parkhill says Paxata enables easy ingestion of diverse data types, easy visualizations to show results and algorithms that permit quick comparisons of patients who have similar genetic makeups and cancer types.
For example, PrecisionProfile is using Paxata’s self-service data prep application, consolidating information that’s stored in Amazon S3, local files and FTP sites. Identifying similar patients employs Paxata’s “patient like this one” algorithm, which enables an oncologist to match a patient to hundreds or even thousands of previously treated patients, making it easier to find similar patients and, more importantly, their response to treatments. Easier access to known treatments and outcomes for genetically, phenotypically and metabolically similar patients provides greater insight and confidence on treatment selection, improving the odds for successful outcomes.
PrecisionProfile contends that it’s able to reduce the cycle time of a genome clinical study from one to three months, to two to eight hours. Additionally, Parkhill says it can cut the time from physician diagnosis to treatment from two to three weeks to a matter of hours.
The ability to bring together public and private genomic data, information from a patient’s electronic health record and clinical records from outside sources lets clinicians focus on researching treatment options, not on managing data, Parkhill contends.
“Our goal is to serve both researchers looking for new drugs or ways to treat new mutations as well as the practicing oncologist,” he adds. While only a few targeted therapies for cancer have been approved by professional organizations, “there are about 750 that are in the pipeline. We can save the oncologist or tumor board a lot of time in figuring out what the right treatment is for a patient.”
Cancers represent different types of genetic mutations of human cells, and some treatments are effective in treating specific types, while others are not; narrowing the search for an effective treatment is crucial in treating cancer, where time is of the essence and survival might hinge on whether clinicians waste time using an ineffective treatment approach.
Although it’s early in the development process, PrecisionProfile plans to provide cloud-based services, although it’s also able to do an installation in a provider’s data center, and it’s able to integrate data from a providers EHR system.
PrecisionProfile is designing a data normalization effort that intends to take those various types of data and speed the process of treating patients. The company expects its efforts will remove some of the drudgery that clinicians now face, facilitate they research they need to do and speed treatment for cancer patients.
Effectively treating these patients requires oncologists to know more than simply in which organ the cancer has its origins. Now, optimizing treatment requires them to know the specific genomic information that precipitated the disease.
“Clinicians now are at the point that the question is now what is the underlying genomic mutation describing the cancer,” says Dave Parkhill, CEO of PrecisionProfile. Then, knowing that, the clinicians may want to use that knowledge to determine what is currently known to be the most effective treatment for patients with that genetic profile for that specific variant of cancer.
Also See: How precision oncology will use data to advance cancer treatmentDelivering those insights is difficult because a patient’s datasets can be scattered in a variety of locations and in several different types of formats. Those researching treatments “are instead wasting precious hours preparing and massaging the data before analysis can even be conducted,” Parkhill says.
For example, cancer genomic data—comprised of a molecular profile of patient cells and chromosomes—is in VCF files and includes a long list of tags and genomic variants for each patient, and the file size can be enormous, with one person’s DNA taking up as much as 50 gigabytes. Demographic and clinical data is typically in XML format, while genetics data can be in diverse formats, often varying depending on the type of cancer.
PrecisionProfile aims to speed the treatment research process by enabling researchers to upload patient-specific information, normalize variances in information, then be able to match patients with similar genetic characteristics and cancers to find specific drugs and treatment protocols that are the most effective.
The company is using a product from Paxata to support the analytics platform. Parkhill says Paxata enables easy ingestion of diverse data types, easy visualizations to show results and algorithms that permit quick comparisons of patients who have similar genetic makeups and cancer types.
For example, PrecisionProfile is using Paxata’s self-service data prep application, consolidating information that’s stored in Amazon S3, local files and FTP sites. Identifying similar patients employs Paxata’s “patient like this one” algorithm, which enables an oncologist to match a patient to hundreds or even thousands of previously treated patients, making it easier to find similar patients and, more importantly, their response to treatments. Easier access to known treatments and outcomes for genetically, phenotypically and metabolically similar patients provides greater insight and confidence on treatment selection, improving the odds for successful outcomes.
PrecisionProfile contends that it’s able to reduce the cycle time of a genome clinical study from one to three months, to two to eight hours. Additionally, Parkhill says it can cut the time from physician diagnosis to treatment from two to three weeks to a matter of hours.
The ability to bring together public and private genomic data, information from a patient’s electronic health record and clinical records from outside sources lets clinicians focus on researching treatment options, not on managing data, Parkhill contends.
“Our goal is to serve both researchers looking for new drugs or ways to treat new mutations as well as the practicing oncologist,” he adds. While only a few targeted therapies for cancer have been approved by professional organizations, “there are about 750 that are in the pipeline. We can save the oncologist or tumor board a lot of time in figuring out what the right treatment is for a patient.”
Cancers represent different types of genetic mutations of human cells, and some treatments are effective in treating specific types, while others are not; narrowing the search for an effective treatment is crucial in treating cancer, where time is of the essence and survival might hinge on whether clinicians waste time using an ineffective treatment approach.
Although it’s early in the development process, PrecisionProfile plans to provide cloud-based services, although it’s also able to do an installation in a provider’s data center, and it’s able to integrate data from a providers EHR system.
More for you
Loading data for hdm_tax_topic #better-outcomes...