Donald P. Harrington

Parameter estimation and tissue segmentation from multispectral MR images

A statistical method is developed to classify tissue types and to segment the corresponding tissue regions from relaxation time T(1 ), T(2), and proton density P(D) weighted magnetic resonance images. The method assumes that the distribution of image intensities associated with each tissue type can be expressed as a multivariate likelihood function of three weighted signal intensity values (T(1), T(2), P(D)) at each location within that tissue regions. The method further assumes that the underlying tissue regions are piecewise contiguous and can be characterized by a Markov random field prior. In classifying the tissue types, the method models the likelihood of realizing the images as a finite multivariate-mixture function. The class parameters associated with the tissue types (i.e. the weighted intensity means, variances and correlation coefficients of the multivariate function, as well as the number of voxels within regions of the tissue types of are estimated by maximum likelihood. The estimation fits the class parameters to the image data via the expectation-maximization algorithm. The number of classes associated with the tissue types is determined by the information criterion of minimum description length. The method segments the tissue regions, given the estimated class parameters, by maximum a posteriori probability. The prior is constructed by the tissue-region membership of the first- and second-order neighborhood. The method is tested by a few sets of T(1), T(2), and P(D) weighted images of the brain acquired with a 1.5 Tesla whole body scanner. The number of classes and the associated class parameters are automatically estimated. The regions of different brain tissues are satisfactorily segmented.

Patency Results of Percutaneous and Surgical Revascularization for Femoropopliteal Arterial Disease

To estimate the patency results of percutaneous transluminal angioplasty and bypass surgery in the treatment of femoropopliteal arterial disease, a Medlars search of the English-language medical literature was performed. Inclusion required that studies 1) report original data, 2) report patency with a life table or Kaplan-Meier analysis with the number at risk or standard errors, 3) define patency as hemodynamic improvement, 4) report the distribution of co variates, and 5) not duplicate other published material. Using a method based on the pro portional-hazards model and the actuarial life-table approach, the results were adjusted for differences in case-mix of the study populations and patency was predicted for subgroups at various levels of risk for failure. The unadjusted pooled life tables yielded five-year pa tencies of 45% (± 2%) for angioplasty, 73% (± 2%) for bypass surgery using a vein graft, and 49% ( ± 3%) for bypass surgery using a polytetrafluoroethylene graft. Adjusted five- year primary patencies after angioplasty varied from 12% to 68%, the best results being for patients with claudication and stenotic lesions. Adjusted five-year primary patencies after surgery varied from 33% to 80%, the best results being for saphenous vein bypass performed for claudication. The authors conclude that pooling life-table data without adjustment for covariates can be misleading. Indication, lesion type, vein graft availability, and site of the distal graft anastomosis need to be considered in predicting patency results of revascular ization for femoropopliteal arterial disease. Key words: arteries, femoral; arteries, popliteal; arteries, transluminal angioplasty; arteries, surgery; review; meta-analysis. (Med Decis Mak ing 1994;14:71-81)

Observations on the use of thrombolytic agents for thrombotic occlusion of infrainguinal vein grafts

Vein graft failure remains a major challenge for the vascular surgeon. Thrombolysis of occluded vein grafts has shown promising short-term results in restoring vein graft patency, however, the long-term results are not established. This study examines the long-term patency and limb salvage after successful thrombolysis and revision of 22 thrombosed vein grafts in 21 patients. There were 17 men and four women with an average age of 60 years (38 to 77 years). Failed vein grafts had an average primary patency of 19 months (1 to 84 months) and included eight in situ grafts and 14 non-in situ grafts. Twelve grafts were to the popliteal level, whereas 10 were infrapopliteal. Thrombolytic agents used included urokinase (15), tissue plasminogen activator (5), and streptokinase (2). After successful thrombolysis, 19 grafts underwent 26 additional procedures including percutaneous transluminal angioplasty (9), vein patch angioplasty (4), vein interposition or jump extension graft (9), or other procedures (4). Three patients had no additional procedure, but one was placed on sodium warfarin (Coumadin). After successful initial vein graft salvage, life-table analysis revealed a 36.6% +/- 11.9% patency at 1 year and a 22.9% +/- 11.6% patency at 3 years. After secondary failure six patients had further interventions contributing to an improved limb salvage of 66.9% +/- 11.6% at 1 year and 60.3% +/- 19.0% at 3 years. The results suggest that thrombosed vein grafts initially salvaged with thrombolysis and revision do not have a favorable long-term patency, and that a premium must be placed on the detection of the failing vein graft before thrombosis.

Deletions of chromosome 13 in multiple myeloma identified by interphase FISH usually denote large deletions of the q arm or monosomy

Deletions of the long arm of chromosome 13 (13q−) are observed in patients with multiple myeloma (MM), are rarely observed in the monoclonal gammopathy of undetermined significance (MGUS) and have been associated with a worsened prognosis in MM. However, no minimally deleted region in the13q arm has been defined at 13q, and consequently no tumor suppressor genes have yet been identified that are important for disease pathogenesis. We attempted to characterize these chromosome 13q deletions at the molecular cytogenetic level. We studied 351 newly diagnosed patients, entered into the E9486/E9487 clinical study of the Eastern Cooperative Oncology Group. Fluorescent in situ hybridization (FISH) combined with immune fluorescent detection (cIg-FISH) of clonal plasma cells (PC) and cytomorphology were used to analyze interphase, bone marrow (BM) cell, cytospin slides. We simultaneously used DNA probes for the following locus specific probes (LSI); LSI 13 (Rb) and D13S319, which hybridize to 13q14. We subsequently studied distal deletions using the D13S25 probe (13q14.3) and a subtelomeric probe (13qSTP) for the 13q-arm (D13S327) in 40 cases with documented LSI 13 (Rb)/D13S319 deletion and 40 without deletion of these loci. Of 325 evaluable patients, we found 13q deletions in 176 (54%) using LSI 13 (Rb) and D13S319 probes. Of 40 patients with LSI 13 (Rb)/D13S319 deletions, 34 (85%) had coexistent deletion of both D13S25/13qSTP. These results indicate that chromosome 13 deletions in MM involve loss of most if not all of the 13q arm perhaps even indicating monosomy. In six cases the 13qSTP signal was conserved, but D13S25 was lost indicating large interstitial deletions involving 13q14. In 39 of the 40 cases without LSI 13 (Rb)/D13S319 deletions, the normal pattern of two pairs of signals was observed for D13S25/13qSTP. Deletions involving 13q14 are very common in MM as detected by cIg-FISH. These deletions appear to predominantly involve loss of large segments of the 13q arm or monosomy 13, and only occasionally represent an interstitial deletion.

Salvage of occluded arterial bypass grafts by means of thrombolysis

Seventy-two thrombosed peripheral arterial bypass grafts in 62 patients were treated by local intraarterial thrombolytic infusion. The initial success rate was 69% (50 of 72 grafts). Graft material and location had no significant effect on the initial results. Urokinase was used in 43 cases with a 84% success rate, and streptokinase was used in 29 cases with a 48% success rate. After a follow-up period that ranged from 2 to 58 months, 27 grafts remained patent, with an average patency duration of 15 months (median 8 months). Overall graft patency at the end of 1 year was 60% applying life-table analysis. Factors that were evaluated to determine their effect on long-term patency included graft age and material, graft location, and the presence or absence of an underlying correctable lesion. The most significant factor in long-term patency was the presence of a lesion that was correctable by surgical revision or balloon angioplasty. In 25 grafts with underlying stenotic lesions, the 1-year patency was 86% after successful treatment. Twenty-five grafts without detectable lesions had 37% 1-year patency.

Risks and benefits of femoropopliteal percutaneous balloon angioplasty

PURPOSE The purpose of this study was to evaluate the efficacy of angioplasty in the treatment of femoropopliteal arterial disease. METHODS From 1980 to 1991, 126 angioplasty procedures were performed in 131 limbs of 106 patients with 175 femoropopliteal lesions (26 common femoral, 118 superficial femoral, and 31 popliteal). Critical ischemia was present in 55 limbs (42%), and claudication was present in 76 (58%). Angioplasty was performed for a single lesion in 87 limbs (66%) and for multiple lesions in 44 (34%). In 13 limbs (10%) the most severe lesion was an occlusion; in 118 (90%) all lesions were stenoses. Distal runoff was good (2 or 3 vessels patent) in 72 limbs (55%) and poor (0 or 1 vessel patent) in 59 (45%). RESULTS Death within 30 days occurred in 0.8%, nonfatal systemic morbidity in 7.1%, and local morbidity in 1.6% of procedures. Multivariate analysis revealed that indication and age were predictive of increased morbidity and mortality rates. Immediate success was achieved in 95% of limbs treated. Mean follow-up time was 2.0 years. The overall 5-year cumulative primary patency rate was 45% (+/- 5%). In a proportional hazards model indication and lesion type were predictive (p < 0.01) of long-term failure, with relative risks of 2.0 (1.2 to 3.3) and 2.7 (1.3 to 5.6), respectively. The 5-year primary patency rate after angioplasty for stenoses and claudication was 55% (+/- 7%), for stenoses and critical ischemia it was 29% (+/- 11%), and for occlusions it was 36% (+/- 14%). CONCLUSION These results suggest that femoropopliteal angioplasty is a low-risk procedure with acceptable long-term results in patients with claudication and stenoses.

Hemodynamic significance of iliac artery stenosis: pressure measurements during angiography.

Peak systolic pressure gradients were obtained before and after vasodilatation in 42 patients (50 limbs) with arteriographic iliac artery stenosis of questionable significance. Patients were divided into three groups according to per cent narrowing of the iliac artery. Pressure gradients across the site of stenosis tended to be significant in patients with greater than 75% stenosis (greater than or equal to 20 mm Hg) but not in patients with less than 50% stenosis; patients in the middle group (50-75% stenosis) demonstrated a wide variance. Thus the arteriogram is not an accurate indicator of hemodynamically significant lesions, particularly in patients with 50-75% stenosis where pressure measurements are of greatest value. Variations in the aortic and femoral artery systolic peak pressure occurred following vasodilatation, indicating the importance of simultaneous pressure recording.

Therapeutic guidelines and results in advanced seminoma.

Twenty patients with advanced seminoma were treated with chemotherapy. Fourteen patients were previously untreated (group 1) and received vinblastine, bleomycin, and cisplatin (VPB) at presentation. Six patients had received prior radiation therapy (group 2), and at relapse received either VPB or VP-16-213 (etoposide)-cisplatin. Within group 1, five patients received no further therapy after VPB (group 1A), six patients received radiation to residual radiographic abnormalities (group 1B), and three patients underwent surgery to remove residual radiographic areas following VPB (group 1C). The complete response rate in group 1 was 14/14 (100%). At present within group 1A, 5/5 patients (100%) are alive and disease-free (NED) for a median follow-up of 32 + months. In group 1B, 6/6 patients (100%) are alive and NED for a median follow-up of 17+ months. In group 1C, 3/3 patients (100%) had residual fibrosis at the time of surgical resection. Two of these patients died of postoperative complications with no evidence of disease and the third is alive and NED at 19+ months. In group 2, 4/6 patients (67%) achieved a complete remission, including two patients who are NED at 22+ and 85+ months, respectively. Two have died and two are alive with progressive disease. Doses of chemotherapy to group 2 patients were substantially lower than the doses given to group 1 patients. We conclude that chemotherapy is acceptable initial therapy for advanced seminoma, and prior extensive radiation therapy may impair the ability to give adequate doses of chemotherapy in patients who relapse. Residual masses after chemotherapy are often fibrotic and the role of postchemotherapy radiation therapy in these patients is uncertain.

Reduction of False Positives by Internal Features for Polyp Detection in CT-Based Virtual Colonoscopy

In this paper, we present a computer-aided detection (CAD) method to extract and use internal features to reduce false positive (FP) rate generated by surface-based measures on the inner colon wall in computed tomographic (CT) colonography. Firstly, a new shape description global curvature, which can provide an overall shape description of the colon wall, is introduced to improve the detection of suspicious patches on the colon wall whose geometrical features are similar to that of the colonic polyps. By a ray-driven edge finder, the volume of each detected patch is extracted as a fitted ellipsoid model. Within the ellipsoid model, CT image density distribution is analyzed. Three types of (geometrical, morphological, and textural) internal features are extracted and applied to eliminate the FPs from the detected patches. The presented CAD method was tested by a total of 153 patient datasets in which 45 patients were found with 61 polyps of sizes 4-30 mm by optical colonoscopy. For a 100% detection sensitivity (on polyps), the presented CAD method had an average FPs of 2.68 per patient dataset and eliminated 93.1% of FPs generated by the surface-based measures. The presented CAD method was also evaluated by different polyp sizes. For polyp sizes of 10-30 mm, the method achieved mean number of FPs per dataset of 2.0 with 100% sensitivity. For polyp sizes of 4-10 mm, the method achieved 3.44 FP per dataset with 100% sensitivity.

Clinical significance of immunophenotype in diffuse aggressive non-Hodgkin's lymphoma.

We performed a prospective study of the clinical significance of immunophenotype in 110 patients with aggressive non-Hodgkins lymphoma (NHL) treated by oncologists in the Nebraska Lymphoma Study Group between October 1982 and May 1986. All patients were immunophenotyped from biopsies performed before therapy was administered. The patients were treated with a uniform protocol of radiotherapy for minimal nonbulky, stage I or II disease (seven patients) or a single, six-drug chemotherapy regimen cyclophosphamide, doxorubicin, procarbazine, bleomycin, vincristine, and prednisone (CAP-BOP) in patients with more extensive disease (103 patients). Ninety-one patients (83%) had B-cell lymphoma and 19 patients (17%) had T-cell lymphoma. The histologic diagnosis of diffuse mixed-cell lymphoma was significantly associated with T-cell immunophenotype (45% v 5%; P less than .001), and the diagnosis of diffuse large-cell lymphoma was significantly associated with B-cell immunophenotype (40% v 5%; P = .006). However, no significant difference in frequency of prognostic variables such as age, stage, systemic symptoms, tumor bulk, serum lactic dehydrogenase, or performance status was found between the B-cell and T-cell groups. Patients with B-cell NHL had a slightly higher complete remission rate (74% v 53%; P = NS), similar durability of complete remission (75% v 70% at 3 years; P = NS), and a slightly but not significantly better overall survival (50% v 41% at 3 years; P = NS). The slight advantage in response rate and survival for B-cell patients was related to a very poor outcome for patients with stage IV T-cell NHL. For patients with stage I to III disease, neither the complete remission rate (B-cell, 82% v T-cell, 91%; P = NS) nor overall survival (3-year survival for B cell, 58% v T cell, 73%; P = NS) were significantly different. However, with stage IV disease B-cell patients fared far better than those with T-cell NHL for both complete remission rate (67% v 0%; P = .002) and overall survival (3-year survival, 44% v 0%; P = .002). Immunophenotyping intermediate- and high-grade NHL allowed identification of a subgroup of patients who had a very poor prognosis with this treatment approach and for whom alternate therapy might be considered.