Each year, nearly 3,000 children in the U.S. are diagnosed with brain tumors. Almost half will die within five years, making it the most fatal cancer among children. If a child does survive a brain tumor, the long-term effects can be significant, including neurological disabilities, retardation and psychological problems.
Beyond surgery, successful treatments for pediatric brain tumors are rare. In fact, clinicians must often conduct experimental research and protocols to develop the treatments. Dr. Eric Bremer, a former director of brain tumor research at Children’s Memorial Research Center in Chicago, was regarded as one of the leading scientists searching for a better way to treat brain tumors. One of his priority projects involved building a gene expression knowledgebase for pediatric brain tumors.
One of the first requisites for this cancer treatment database is accurately classifying the tumor. While the vast majority of pediatric brain tumors can be categorized as gliomas, ependymomas and medulloblastomas, altogether there are 12 or so tumor types, including subtypes. Classification is very subjective because pediatric brain tumors are often difficult to distinguish by appearance, and there are few objective markers such as those found for other childhood cancers like leukemia.
In addition to tumor type, it's important to classify a tumor's stage or grade. Cancers are generally stratified into four stages, from the most benign (stage I) to malignant (stage IV). Neuropathologists often find it difficult to distinguish between intermediate stages. However, treatment between two stages can be drastically different, and incorrect staging can have dramatic consequences for the patient. For example, if a child with a stage two tumor is misdiagnosed with a more advanced grade, he would unnecessarily receive more aggressive treatment. This not only results in needless pain, but also could lead to long-term damage as mentioned earlier.
A key to accurate classification of pediatric brain tumors lies at the molecular level. Just as a skin cell and liver cell vary in their gene expression patterns, the same is true for different tumors or tumor grades. Dr. Bremer was able to capture these differences with gene expression microarray experiments, but analyzing these differences to identify and classify the tumor is a more formidable obstacle. Said Dr. Bremer, "You can easily get 7,000 to 30,000 data points for each sample. The problem is how to make sense of them."
Dr. Bremer literally ran into the solution to his problem at the Microarray Data Analysis Conference when he met Cathy DeSesa and Petra Scheffer of SPSS Inc. Later, DeSesa, a project manager, and Fabrice Leroy, an SPSS data mining solution architect, consulted and worked with Dr. Bremer to analyze his microarray data using SPSS' data mining solution, Clementine, the leading data mining workbench. Traditionally used in customer oriented business intelligence applications, Clementine is now being used in the life sciences to study microarray data. Pre-built streams (a flow of data-mining steps) called Microarray Clementine Application Templates (MCAT) comprise a basic outline for analyzing gene expression data, and represent the best analytical practices in the field.
Interested in Children’s Memorial Hospital? Download the PDF
There is no one right way to analyze the data; you want to try several ways. What's so great about Clementine is that it's a workbench. It's easy to try different methodologies in order to find a method, or combination of methods that works best.
Dr. Eric Bremer
Director of brain tumor research
Children's Memorial Hospital, Chicago
Dr. Bremer was won over by Clementine for one simple reason: "it worked." Dr. Bremer combined his own data with that of the publicly available Pomeroy et. al. data set (Nature vol. 415, pp. 436-442, 2002) resulting in a total of 133 tumor samples from the six major pediatric brain tumor types. Clementine successfully classified these tumors with greater than 95 percent accuracy. While these samples were well described pathologically, they served as a test case that bodes well for future pediatric brain tumor classification, especially difficult-to-classify tumors. In addition, sub-classification of gliomas and medulloblastomas also appeared to be very reliable.
The key was Clementine's flexibility. Prior to using Clementine, Dr. Bremer used up to five different software packages to analyze microarray gene expression data. Clementine replaced all of those packages.
"There is no one right way to analyze the data; you want to try several ways," said Dr Bremer. "What's so great about Clementine is that it's a workbench. It's easy to try different methodologies in order to find a method, or combination of methods that works best."
Dr. Bremer used two of Clementine's predictive models, artificial neural networks and decision trees, to analyze and classify his data. Information from one model complemented the other. While the neural network resulted in a more accurate classification, it didn't show how it actually accomplished the classification. The decision tree, however, showed precisely how the tumors were classified and revealed potential gene markers that characterized certain cancers. Once these markers are validated, labs without microarray technology could use this information to develop antibodies against the gene product (protein) as an alternate means of diagnosis.
Just as important as the model is the data that the model is based on. Microarray data analysis presents a number of challenges given the small number of samples and large number of genes. Dr. Bremer's data set had 133 samples, but nearly 7,000 variables. The Microarray CAT is based on real-world experiences to overcome these challenges.
It's also important that differences in gene expression values genuinely reflect biological variation and not artificial differences introduced during sample preparation. Clementine can help assure this is the case by processing the data quality parameters along with the samples. Clementine also simplifies the task of feeding data into the model. Prior to Clementine, Dr. Bremer had to organize the expression values in a specific format dependent on the requirements of the analysis package. Dr. Bremer now skips this step because Clementine streams can automatically prepare and accept gene expression data directly from his database — a huge time saver.
Classification using data obtained from gene expression microarray experiments is just part of a bigger picture. Expression measures RNA levels and not the final gene products. Dr. Bremer's database will eventually incorporate clinical, pathological and biochemical information to provide as complete and accurate a picture as possible. In addition, patient outcome information will be added so that the database may reveal which treatments work best with brain tumors sharing genetic and pathological characteristics. This is the point at which bioinformatics transition to "biomedical informatics."
Dr. Bremer also realizes the critical need to share data and knowledge. To this end, he plans to deploy Clementine streams via a Web server so that any pediatric brain tumor researcher can submit files from their own microarray experiments for analysis, and then receive predictive responses. Clementine will enable Dr. Bremer's data to be combined with data from other researchers. In turn, the added data can be used to update and add value to the original data set; the more samples his database contains, the more accurate it will be and the more lives that will be saved.
The bigger database will also help reveal a multitude of genes that are involved in tumor development. According to Nature, in 2001 alone, more than 21,000 articles on characterizing, diagnosing and treating malignancies were published. Dr. Bremer will use SPSS LexiQuest® Mine to sift through this literature and extract patterns that, for example, when combined with data from his tumor database, might help him evaluate prime drug targets that would form the basis for a cure for cancer. This, after all, is Dr. Bremer's ultimate goal.
The complete list of global SPSS success stories can be found here
Key Products and Services
Resources
Demos and Downloads
Was this information useful?