[Show abstract][Hide abstract] ABSTRACT: New advances in medicine have led to a disparity between the existing information about patients and the ability of clinicians to utilize it. Lack of training and incompatibility with clinical techniques has made the use of the complex adaptive systems approach difficult. To avoid this, we used statistical learning theory as an inline preprocess between existing data collection methods and clinical analysis of data. Clinicians would be able to use this system without any changes to their techniques, while improving accuracy. We used data from CT scans of patients with metastatic carcinoma to predict prognosis. Specifically, we used the standard for evaluating response to treatment, RECIST, and new qualitative and quantitative features. An Evolutionary Programming trained Support Vector Machine (EP-SVM), was used to preprocess the data for two traditional survival analysis techniques: Cox Proportional Hazard Models and Kaplan Meier curves. This was compared to Logistic Regression (LR) and using cutoff points. Analyses were also done to compare different inputs and different radiologists. The EP-SVM outperformed both LR and the cutoff method significantly and allowed us to both intelligently combine data from multiple sources and identify the most predictive features without necessitating changes in clinical methods.
[Show abstract][Hide abstract] ABSTRACT: While the role of survival analysis in medicine has continued to be increasingly essential in making treatment and other health care decisions, the common clinical methods used for performing these analyses, such as Cox Proportional Hazard models and Kaplan-Meier curves, have become antiquated. We have developed a new survival analysis technique of the Evolutionary Programming / Evolutionary Strategies Support Vector Regression Hybrid for censored and non-censored event data. This method provides the benefits of optimized statistical learning theory to be used as a replacement for or in addition to existing survival analysis protocols. The technique was tested on an artificially censored data from a well-known benchmark dataset as well as actual clinical data with encouraging results.
[Show abstract][Hide abstract] ABSTRACT: To assess whether quantitative visual scoring (QVS) is a better early predictor of progression-free survival (PFS) in patients on chemotherapy for metastatic melanoma using CT than the currently used Response Evaluation Criteria in Solid Tumors (RECIST) standard. Retrospective evaluation of 65 consecutive patients with metastatic melanoma on treatment who had a baseline and follow-up CT after two cycles of therapy. QVS was used to code imaging findings on the radiology reports considering size change, brain metastases, new lesions, mixed lesion response, and the number of organ systems involved. RECIST 1.1 criteria placed patients in the progressive disease, stable disease, or partial response groups. Multiple regression analysis was used to correlate the various independent variables with PFS. The Cox hazard proportions ratio, median survival, and Kaplan-Meier curves of the different prognostic groups were calculated. QVS of size change was found more sensitive in detecting patients deteriorating (57.1% versus 37.5%) or improving (23.8% versus 10.7%), more correlated with the median PFS for the deteriorating (1.8 versus 1.7 months), stable (5.6 versus 4.0 month), and improving (8.3 versus 5.5 months) categories and more predictive of PFS (Cox hazard proportion ratio of 3.070 versus 1.860) than RECIST 1.1 categorization. Multiple regression analysis demonstrated QVS of lesion size correlated most closely with PFS among the variables assessed (r = 0.519, p < 0.0001). QVS in this study was superior to standard RECIST categorization in terms of discriminating treated metastatic melanoma patients likely to have longer PFS.
Journal of Digital Imaging 07/2011; 25(2):258-65. DOI:10.1007/s10278-011-9407-9 · 1.19 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: As cancer treatment cost soar and the mantra for "personalized medicine" grows louder, we will increasingly be searching for solutions to these diametrically opposed forces. In this review we highlight several exciting novel imaging strategies including MRI, CT, PET SPECT, sentinel node, and ultrasound imaging that hold great promise for improving outcomes through detection of lymph node involvement. We provide clinical data that demonstrate how these evolving strategies have the potential to transform treatment paradigms.
[Show abstract][Hide abstract] ABSTRACT: Significant interest exists in establishing radiologic imaging as a valid biomarker for assessing the response of cancer to a variety of treatments. To address this problem, we have chosen to study patients with metastatic colorectal carcinoma to learn whether statistical learning theory can improve the performance of radiologists using CT in predicting patient treatment response to therapy compared with the more traditional RECIST (Response Evaluation Criteria in Solid Tumors) standard.
Predictions of survival after 8 months in 38 patients with metastatic colorectal carcinoma using the Support Vector Machine (SVM) technique improved 30% when using additional information compared to WHO (World Health Organization) or RECIST measurements alone. With both Logistic Regression (LR) and SVM, there was no significant difference in performance between WHO and RECIST. The SVM and LR techniques also demonstrated that one radiologist consistently outperformed another.
This preliminary research study has demonstrated that SLT algorithms, properly used in a clinical setting, have the potential to address questions and criticisms associated with both RECIST and WHO scoring methods. We also propose that tumor heterogeneity, shape, etc. obtained from CT and/or MRI scans be added to the SLT feature vector for processing.
[Show abstract][Hide abstract] ABSTRACT: To establish radiologic imaging as a valid biomarker for assessing the response of cancer to different treatments. We study patients with metastatic colorectal carcinoma to learn whether Statistical Learning Theory (SLT) improves the performance of radiologists using Computer Tomography (CT) in predicting patient treatment response to therapy compared with traditional Response Evaluation Criteria in Solid Tumours (RECIST) standard. Preliminary research demonstrated that SLT algorithms can address questions and criticisms associated with both RECIST and World Health Organization (WHO) scoring methods. We add tumour heterogeneity, shape, etc., obtained from CT or MRI scans the feature vector for processing.
International Journal of Computational Biology and Drug Design 08/2010; 3(1):15-8. DOI:10.1504/IJCBDD.2010.034463
[Show abstract][Hide abstract] ABSTRACT: To develop a methodology which quantifies multiple changing lesion features resulting in an optimized computed tomography (CT) response score (CRS) for prediction of overall survival (OS) in response to treatment for metastatic colorectal carcinoma (MCRC).
This Health Insurance Portability and Accountability Act-compliant, institutional review board-approved retrospective study evaluated multiple changing imaging findings and their correlation with OS with a new methodology comparing the baseline and first post-treatment CT scans in 38 MCRC patients on last-line chemotherapy (cetuximab and irinotecan). Tumor size/enhancement changes and interval development of new lesions were quantified with either Likert-type scales (all parameters) or Response Evaluation Criteria in Solid Tumors (RECIST) (size change only). The most predictive parameters for OS were used to generate the CRS with an overall range of -3 (complete disappearance) to +2 (definite tumor increase). The Cox Hazard Ratio was used to assess prediction of survival. Reader agreement was evaluated by the kappa statistic.
Tumor size was the best predictor of OS using the Likert-type scale or RECIST. The CRS was not improved combining size change with other parameters. Use of the Likert-type scale resulted in predicting OS with a Cox hazard ratio of 1.697 (P=.0004) and good agreement (kappa=0.73, 95% CI=0.41-1.10) between observers with no significant difference using RECIST.
The methodology produces a CRS for MCRC predicting OS resulting from therapy which expands standard RECIST guidelines to allow critical evaluation of multiple additional imaging parameters. Size change alone was found to be the best parameter of those considered in terms of maximizing agreement and prediction of OS.
[Show abstract][Hide abstract] ABSTRACT: Lung cancer is the leading cause of cancer death in the United States. Mortality outcomes have improved only modestly over the past 30 years. There is intense focus on the development of better treatments for lung cancer. Major issues include the cost and time duration of the clinical trials required to establish the utility of a drug so that it can be formally approved by regulatory agencies. In clinical settings, biomarkers that accelerate assessments of responses to treatment could benefit patients by providing earlier diagnoses of progressive disease, particularly when there are multiple options for treatment, and the effects of toxicity from one treatment tend to limit the ability to administer the next line of therapy.
Quantifying longitudinal changes in tumor volumes using computed tomography could eventually become a more useful surrogate endpoint for assessing tumor responses or progression events than simple unidimensional measurements.
The authors review the historical development of response measurements in lung cancer, set out the medical context for specifying volumetric imaging requirements and goals, compare volumetric technique to conventional methods, and identify the imaging profiles being pursued.
The Quantitative Imaging Biomarkers Alliance is investigating volumetric computed tomographic acquisition and analytic methods to increase the analytic power per subject enrolled in clinical trials to reduce the number of total subjects needed or shorten the length of time an individual needs to be followed to reliably establish drug response.
[Show abstract][Hide abstract] ABSTRACT: Our objective was to compare a newly developed semiquantitative visual scoring (SVS) method with the current standard, the Response Evaluation Criteria in Solid Tumors (RECIST) method, in the categorization of treatment response and reader agreement for patients with metastatic lung cancer followed by computed tomography.
The 18 subjects (5 women and 13 men; mean age, 62.8 years) were from an institutional review board-approved phase 2 study that evaluated a second-line chemotherapy regimen for metastatic (stages III and IV) non-small cell lung cancer. Four radiologists, blinded to the patient outcome and each other's reads, evaluated the change in the patients' tumor burden from the baseline to the first restaging computed tomographic scan using either the RECIST or the SVS method. We compared the numbers of patients placed into the partial response, the stable disease (SD), and the progressive disease (PD) categories (Fisher exact test) and observer agreement (kappa statistic).
Requiring the concordance of 3 of the 4 readers resulted in the RECIST placing 17 (100%) of 17 patients in the SD category compared with the SVS placing 9 (60%) of 15 patients in the partial response, 5 (33%) of the 15 patients in the SD, and 1 (6.7%) of the 15 patients in the PD categories (P < 0.0001). Interobserver agreement was higher among the readers using the SVS method (kappa, 0.54; P < 0.0001) compared with that of the readers using the RECIST method (kappa, -0.01; P = 0.5378).
Using the SVS method, the readers more finely discriminated between the patient response categories with superior agreement compared with the RECIST method, which could potentially result in large differences in early treatment decisions for advanced lung cancer.
[Show abstract][Hide abstract] ABSTRACT: Multiple myeloma (MM) remains an incurable cancer. Treatment often is initiated at the time patients experience a progressive increase in tumor burden. The authors of this report investigated magnetic resonance imaging of the bone marrow (BM-MRI) as a novel approach to quantify disease burden and validated a staging system by correlating BM-MRI with common clinical and laboratory parameters.
The extent of bone marrow involvement was evaluated by BM-MRI. Clinical and laboratory parameters were assessed in patients with active MM, and correlations between variables were assessed statistically. Bone marrow involvement by BM-MRI was defined as stage A (0%), stage B (<10%), stage C (10%-50%), and stage D (>50%).
In total, 170 consecutive patients were evaluated (77 women and 93 men), including 144 patients who had active MM. The median age was 61 years (age range, 35-83 years). Advance stage disease (stage >I) based on Durie-Salmon (DS) staging or International Staging System (ISS) criteria was observed in 122 patients (84%) and 77 patients (53%), respectively. Lytic bone disease was noted in 120 patients (83%). There was a significant association between BM-MRI involvement and DS stage (P = .0006), ISS stage (P = .0001), the presence of lytic bone disease (P < .0001) and mean beta-2 microglobulin levels (P < .0001). Among the patients with previously untreated MM, there was a significant association between BM-MRI stage and overall survival (OS) (univariate P = .013; multivariate P = .045). Plasmacytosis on bone marrow biopsy at diagnosis was not predictive of OS (P = .91).
BM-MRI is a novel approach for quantifying disease burden in patients with MM. The current investigation in a large cohort of nontransplantion MM patients demonstrated that the extent of bone marrow involvement determined by BM-MRI correlates accurately with other conventional parameters of disease burden and can independently predict survival in patients with MM at the time of initial diagnosis.
Cancer 01/2009; 116(1):84-92. DOI:10.1002/cncr.24704 · 4.89 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: We had previously demonstrated that low dose irinotecan (CPT-11) leads to increased accumulation of cells in S-phase and shows a therapeutic synergy with S-phase specific chemotherapy such as gemcitabine and 5-fluorouracil. In this phase II study, our objectives were to evaluate the tolerability and activity of low dose CPT-11 followed 24 h later by gemcitabine as second line therapy in patients with metastatic non-small cell lung cancer (NSCLC). CPT-11 (60 mg/m) was administered 24 h before gemcitabine (1000 mg/m) on days 1, 2, 8, and 9 every 3 weeks. Twenty-nine patients were evaluable for response. The median follow-up was 7.4 months. Partial response (PR) was seen in two (6.9, 95% confidence interval (CI): 0.009-0.228). PR and stable disease were seen in 22 patients (75.9, 95% CI: 0.564-0.897). The median survival time was 13.8 months (95% CI: 8.1-19.3). The median time to progression was 4.6 months (95% CI: 2.6-6.2). Thirty-eight patients were evaluable for toxicity. Neutropenia (grade 3 or 4) was observed in 27 patients (71%). Eight patients did not receive cycle 2 of therapy owing to prolonged neutropenia. No treatment-related deaths occurred. Scheduled administration of low dose CPT-11, 24 h before gemcitabine in the second line therapy of NSCLC yielded comparable disease control rates (PR and stable disease) when compared with other studies using the two chemotherapy drugs in the traditional sequence. However, this approach was associated with higher grade 3/4 neutropenia and is not recommended for further study in metastatic NSCLC.
[Show abstract][Hide abstract] ABSTRACT: Our purpose was to assess whether a simpler qualitative evaluation of tumor response by computed tomography is as reproducible and predictive of clinical outcome as the Response Evaluation Criteria in Solid Tumors (RECIST) and World Health Organization (WHO) methods. This study was a two-reader retrospective evaluation in which qualitative assessment resulted in agreement in 21 of 23 patients with metastatic colorectal carcinoma (91.3%, kappa=0.78; 95% CI, 0.51-1.00). Hepatic metastases were classified as increased, decreased, or unchanged, compared with agreement in 20 of 23 patients (87.0%) for RECIST (kappa=0.62; 95% CI, 0.23-1.00) and WHO (kappa=0.67; 95% CI, 0.34-1.00) methods. Patients were placed into partial response, stable disease, and disease progression categories. Time to progression of disease was better predicted qualitatively than by RECIST or WHO. Our pilot data suggest that our qualitative scoring system is more reproducible and predictive of patient clinical outcome than the RECIST and WHO methods.
[Show abstract][Hide abstract] ABSTRACT: To assess the risk of contrast-induced nephropathy in cancer patients with underlying renal insufficiency receiving the iso-osmolar intravenous contrast agent iodixanol for diagnostic computed tomography (CT) examinations.
Institutional review board approval was obtained with waiver of informed consent. Our study was a retrospective evaluation comparing the incidence of contrast-induced nephropathy in consecutive patients with underlying renal insufficiency undergoing diagnostic CT examinations receiving iodixanol from November 2003 to June 2005 with a comparison group of patients with normal baseline renal function over the same period. Renal insufficiency was considered a serum creatinine level more than 1.2 mg/dL in females and more than 1.5 mg/dL in males. Contrast nephropathy was considered an absolute elevation of 0.5 mg/dL or 25% elevation in serum creatinine level.
In the group of patients receiving iodixanol with underlying renal insufficiency (189 patients), 9.0% developed contrast nephropathy (P = 0.015) with 4.8% of patients developing irreversible renal damage (P = 0.03). This compared with 4.9% of patients receiving iodixanol (185 patients) and 3.1% of patients receiving iohexol (194 patients) with normal baseline renal function developing contrast nephropathy (P = 0.38) with 3.2% of the iodixanol patients and 1.0% of the iohexol patients developing irreversible renal damage (P = 0.13).
The risk of contrast-induced nephropathy is significantly higher in patients with underlying renal insufficiency receiving iodixanol than that for patients with normal baseline renal function, but this should not serve as an absolute contraindication for these patients to receive intravenous iodinated contrast for diagnostic CT examinations particularly in patients with life-threatening clinical questions in which contrasted CT may provide valuable information.
[Show abstract][Hide abstract] ABSTRACT: PURPOSE
To assess the risk of contrast induced nephropathy in cancer patients with underlying renal insufficiency receiving the iso-osmolar intravenous contrast agent iodixanol for diagnostic CT examinations.
METHOD AND MATERIALS
Retrospective evaluation comparing the incidence of contrast induced nephropathy in consecutive patients undergoing diagnostic CT examinations receiving iodixanol from November 2003 to June 2005 with a randomly selected group receiving iohexol with normal baseline renal function over the same time period. Renal insufficiency was considered a serum creatinine level > 1.2 mg/dL in females and > 1.5 mg/dL in males. Contrast nephropathy was considered an absolute elevation of 0.5 mg/dL or 25% elevation in serum creatinine level.
In the group of patients receiving iodixanol with underlying renal insufficiency (190 patients), 8.9% developed contrast nephropathy (p=0.02) with 4.7% of patients developing permanent renal damage (p=0.03). This compared with 4.9% of patients receiving iodixanol (185 patients) and 3.1% of patients receiving iohexol (194 patients) with normal baseline renal function developing contrast nephropathy (p=0.38) with 3.2% of the iodixanol patients and 1.0% of the iohexol patients developing permanent renal damage (p=0.13).
The risk of contrast induced nephropathy is slightly higher in patients with underlying renal insufficiency receiving iodixanol than that for patients with normal baseline renal function, but this should not serve as an absolute contraindication for these patients to receive IV contrast for diagnostic CT examinations particularly in patients with life threatening clinical questions in which contrasted CT may provide valuable information.
This study provides the risk of contrast induced nephropathy in patients with renal insufficiency receiving the iso-osmolar contrast agent iodixanol for diagnostic CT examinations.
Radiological Society of North America 2006 Scientific Assembly and Annual Meeting; 11/2006
[Show abstract][Hide abstract] ABSTRACT: OBJECTIVE: Radiologic imaging examinations are being ordered beyond the margin of medical necessity. Radiologists can assess the value of imaging in a variety of clinical situations by gathering data regarding test ordering patterns and their effects on patient outcomes. CONCLUSION: Emerging information technologies have the potential to facilitate the collection of data and permit the dissemination of appropriate guidelines to limit the number of unnecessary and possibly harmful examinations.
American Journal of Roentgenology 01/2006; 185(6):1399-403. DOI:10.2214/AJR.05.0482 · 2.73 Impact Factor