Release date: 2016-10-31
What is most important to you in life?
Different people have very different answers. For MIT's mathematician Professor Dimitris Bertsimas, the most important thing in his life is three things: improving the human condition; bringing positive impact on people's lives; understanding this How the world works.
Dimitris is a well-known mathematician at MIT
When writing his philosophy of affairs on the MIT official website, Dimitris may not have thought of what important changes he would have for cancer patients over the years.
The cause of the story happened nearly 10 years ago. In 2007, Dimitris's father was diagnosed with non-metastatic gastric cancer, and surgery was powerless. The only treatment option was chemotherapy. In order to let his father prolong life and improve his quality of life, Dimitris studied the conventional chemotherapy regimen of the five major hospitals in the United States. However, after some investigations, he accidentally discovered that the strategy used by each hospital is different. Which treatment method can have a better effect on my father?
“I read the clinical trial data from these hospitals and did some simple calculations,†Dimitris said. The nature of the mathematician allowed him to think about different chemotherapy strategies from a rational perspective - he drew a picture, the abscissa is the toxicity of the drug, and the ordinate is the survival rate of the patient. On the curve, Dimitris took one of his best strategies and asked his father to start treatment.
His father lived for two more years after his diagnosis, which is double the doctor's expectations.
Although the incidence of gastric cancer is still decreasing, the 5-year survival rate of patients with gastric cancer is still not optimistic.
Although he did not eventually save his father's life, this experience gave Dimitris a new revelation - if more time is spent collecting more data, can more patients benefit from it? So, he and his research team collected a large number of papers on gastric cancer since 1980, and obtained useful clinical trial data from 414 of them. “We got critical information from a large number of patients, including how serious they were, what drugs they used, how they were treated and how effective they were, how many side effects they had, and how long they lived,†Dimitris Says: "These data can tell us which treatments work best for specific cancers and people, and which ones have side effects."
Using this data, the research team built a statistical model using machine learning. In this model, the type of drug and their dose are directly linked to the final effect. In other words, as long as the drug information is filled in this model, this model can theoretically predict the effect of chemotherapy.
"We know that this model is particularly useful," Dimitris said. The research team used data from 1980-2005 and 1980-2006 to predict the results of the trials in 2006 and 2007. The data returned by the computer shows that the model's predictions are extremely accurate.
The prediction results of the existing data give the research team a great confidence in the accuracy of the model.
This exciting result means that we can understand the possible effects of different combinations of chemotherapy drugs before the start of clinical trials. Dimitris said that we may have tried the combination therapy of drug A and drug B before, and also tried the combination therapy of drug C and drug D, but the therapy of drug A and drug C is a brand new field, no one knows them. What is the effect? Now, we are able to find a combination of drugs that can best serve a specific population from a large drug database for better treatment.
This has achieved miraculous effects.
Taking advanced gastric cancer as an example, after the diagnosis of this malignant tumor, the average life expectancy of the patient is only about 10 months. However, after treatment with an optimized combination of drugs, the patient's life span can be extended by an additional 4-6 months. This is equivalent to an increase of 40%-60% of life. "We can't cure cancer at the moment, but we can extend the patient's survival, reduce the toxicity of chemotherapy, and let patients live a more quality life," Dimitris said.
This is not the whole of this powerful model. The researchers found that it is expected to find patients who are not sensitive to specific therapies before the trial. This allows the clinical trial designer to reconsider the patient's recruitment conditions, streamline the conditions, and save the cost of the trial. In addition, with the support of subsequent tests, these results may lead to new biomarkers, or to find new mechanisms of disease, further affecting people's understanding of the disease.
The research team expects to apply this model to more diseases.
Currently, the research team is applying the same approach to other diseases outside of gastric cancer. Recently, they have just completed data collection of more than 1,500 clinical trials of breast cancer chemotherapy, and data analysis and modeling are also in full swing. It is reported that they plan to cooperate with the New York Presbyterian Hospital to promote more precise clinical trials. In addition, Dimitris has also begun to study the specific efficacy of diabetes drugs, and hope to bring personalized glycemic control programs to patients. Ultimately, the research team plans to work with the American Society of Clinical Oncology to make all the data and models public.
In this way, the efforts of a mathematician to extend his father's life ultimately have a positive impact on the majority of cancer patients. And this is what Dimitris is most valued. We wish this accurate model combined with machine learning to be open soon, and to make the most effective treatments for the world's major diseases, including Chinese patients.
Source: WuXi PharmaTech
YokeLink Drop Wire Clamp is used to support one and two pair telephone or fiber drop wire at span clamps, drive hooks, house hooks and various drop attachments. Made from 3 pieces stainsteel removable construction, Dimpled, Serrated and Plastic shim are available.
YokeLink Cable Lashing Clamp used for securing lashing wire at end of span.
YokeLink Acetal Strap Cable Ties made by toughened weather acetal. Temperature Range:
-40℃ to 85℃. Black colour, general packing in 50 feet length, can cut the strap at different length in practice. The Double Locking Heads and Tool are available.
YokeLink Stainless steel Banding have high acid resistance, corrosion resistance and tensile strength, easy and convenient operation.
Aerial Drop Hardware,Cable Span Clamp,Tooth Type Band Buckle,Acetal Strap Locking Head
Ningbo Yokelink Machinery Co.,Limited , https://www.yokelink.com