"Time Saved" Calculations to Improve Decision-Making in Progressive Disease Studies.

BACKGROUND: Disease modifying therapies (DMTs) may be most beneficial in early disease, when progression is slow and changes small, with clinical relevance difficult to interpret. OBJECTIVES: Time component tests (TCTs) translate differences between treatments from mean change, vertical distance between longitudinal trajectories, into intuitively understood time saved, horizontal distance between trajectories, which can be readily combined across endpoints in a global TCT (gTCT). DESIGN: The value of composites, time savings estimates, and combination scores to optimize measurement and interpretation of DMTs are demonstrated, along with construction details and simulation studies. SETTING: TCT methods were applied to a randomized phase II clinical trial. PARTICIPANTS: Patients with early Alzheimer's disease (N=332). INTERVENTION: Three treatment groups with AFFITOPE® AD02 and two control groups with aluminum oxyhydroxide, AD04. MEASUREMENTS: The co-primary efficacy outcomes were an adapted ADAS-Cog (aADAS) and adapted ADCS-ADL (aADL), which were optimized composite scales specific to cognitive and functional domains. A composite based on these two scores was the study's prespecified primary outcome. The CDR-sb and standard non-adapted ADCS-ADL and ADAS-Cog scales were prespecified secondary outcomes. RESULTS: The AD04 2 mg group showed some statistically significant effects compared with other study arms. It is unclear whether the observed 3.8-point difference on the composite is clinically meaningful. TCT results show a time savings of 11 months in an 18-month study with AD04 2 mg. CONCLUSION: The relevance of 11 months saved is more universally understood than a mean difference of 3.8 points in the composite outcome. These results suggest that a combination of a composite approach and a time savings interpretation offers a powerful approach for detecting and interpreting disease modifying effects.

Aligning Estimators With Estimands in Clinical Trials: Putting the ICH E9(R1) Guidelines Into Practice.

The draft ICH E9(R1) addendum stipulates that an estimator should align with its associated estimand and yield an estimate that facilitates reliable interpretations. The addendum further stipulates that assumptions should be justifiable and plausible, and that the extent of assumptions is an important consideration for whether an estimate will be robust because assumptions are often unverifiable. The draft addendum specifies 5 strategies for dealing with intercurrent events. The intent of this paper is to provide conceptual considerations and technical details for various estimators that align with these strategies. We include focus on how the nature and extent of assumptions influences the potential robustness of the various estimators. The content reflects the knowledge, experience, and opinions of the Drug Information Association's Scientific Working Group on Missing Data. This group includes experienced statisticians from across industry and academia, primarily in the US and European Union.

A structured framework for assessing sensitivity to missing data assumptions in longitudinal clinical trials.

The objective of this research was to demonstrate a framework for drawing inference from sensitivity analyses of incomplete longitudinal clinical trial data via a re-analysis of data from a confirmatory clinical trial in depression. A likelihood-based approach that assumed missing at random (MAR) was the primary analysis. Robustness to departure from MAR was assessed by comparing the primary result to those from a series of analyses that employed varying missing not at random (MNAR) assumptions (selection models, pattern mixture models and shared parameter models) and to MAR methods that used inclusive models. The key sensitivity analysis used multiple imputation assuming that after dropout the trajectory of drug-treated patients was that of placebo treated patients with a similar outcome history (placebo multiple imputation). This result was used as the worst reasonable case to define the lower limit of plausible values for the treatment contrast. The endpoint contrast from the primary analysis was - 2.79 (p = .013). In placebo multiple imputation, the result was - 2.17. Results from the other sensitivity analyses ranged from - 2.21 to - 3.87 and were symmetrically distributed around the primary result. Hence, no clear evidence of bias from missing not at random data was found. In the worst reasonable case scenario, the treatment effect was 80% of the magnitude of the primary result. Therefore, it was concluded that a treatment effect existed. The structured sensitivity framework of using a worst reasonable case result based on a controlled imputation approach with transparent and debatable assumptions supplemented a series of plausible alternative models under varying assumptions was useful in this specific situation and holds promise as a generally useful framework.

A structured approach to choosing estimands and estimators in longitudinal clinical trials.

An important evolution in the missing data arena has been the recognition of need for clarity in objectives. The objectives of primary focus in clinical trials can often be categorized as assessing efficacy or effectiveness. The present investigation illustrated a structured framework for choosing estimands and estimators when testing investigational drugs to treat the symptoms of chronic illnesses. Key issues were discussed and illustrated using a reanalysis of the confirmatory trials from a new drug application in depression. The primary analysis used a likelihood-based approach to assess efficacy: mean change to the planned endpoint of the trial assuming patients stayed on drug. Secondarily, effectiveness was assessed using a multiple imputation approach. The imputation model-derived solely from the placebo group-was used to impute missing values for both the drug and placebo groups. Therefore, this so-called placebo multiple imputation (a.k.a. controlled imputation) approach assumed patients had reduced benefit from the drug after discontinuing it. Results from the example data provided clear evidence of efficacy for the experimental drug and characterized its effectiveness. Data after discontinuation of study medication were not required for these analyses. Given the idiosyncratic nature of drug development, no estimand or approach is universally appropriate. However, the general practice of pairing efficacy and effectiveness estimands may often be useful in understanding the overall risks and benefits of a drug. Controlled imputation approaches, such as placebo multiple imputation, can be a flexible and transparent framework for formulating primary analyses of effectiveness estimands and sensitivity analyses for efficacy estimands.

The enrichment window approach as a means of dealing with placebo response in antidepressant clinical trials.

In this issue, Merlo-Pich et al. present an enrichment-window approach that identifies sites with aberrant mean placebo responses and excludes data from those sites so as to improve drug-placebo discrimination in antidepressant clinical trials. The method appears to increase the signal in situations in which the test drug is better than placebo. However, confirmation of its impact on the rate of false-positive results is needed before the method can be used prospectively.

The impact of analytic method on interpretation of outcomes in longitudinal clinical trials.

AIMS: Various analytical strategies for addressing missing data in clinical trials are utilised in reporting study results. The most commonly used analytical methods include the last observation carried forward (LOCF), observed case (OC) and the mixed model for repeated measures (MMRM). Each method requires certain assumptions regarding the characteristics of the missing data. If the assumptions for any particular method are not valid, results from that method can be biased. Results based on these different analytical methods can, therefore, be inconsistent, thereby making interpretation of clinical study results confusing. In this investigation, we compare results from MMRM, LOCF and OC in order to illustrate the potential biases and problems in interpretation. METHODS: Data from an 8-month, double-blind, randomised, placebo-controlled (placebo; n = 137), outpatient depression clinical trial comparing a serotonin-noradrenalin reuptake inhibitor (SNRI; n = 273) with a selective serotonin reuptake inhibitor (SSRI; n = 274) were used. The study visit schedule included efficacy and safety assessments weekly to week 4, bi-weekly to week 8, and then monthly. Visitwise mean changes for the 17-item Hamilton Depression Rating Scale (HAMD(17)) Maier subscale (primary efficacy outcome), blood pressure, and body weight were analysed using LOCF, MMRM and OC. RESULTS: Last observation carried forward consistently underestimated within-group mean changes in efficacy (benefit) and safety (risk) for both drugs compared with MMRM, whereas OC tended to overestimate within-group changes. CONCLUSIONS: Inferences are based on between-group comparisons. Therefore, whether or not underestimating (overestimating) within-group changes was conservative or anticonservative depended on the relative magnitude of the bias in each treatment and on whether within-group changes represented improvement or worsening. Preference should be given in analytic plans to methods whose assumptions are more likely to be valid rather than relying on a method based on the hope that its results, if biased, will be conservative.

Patterns of depressive symptom response in duloxetine-treated outpatients with mild, moderate or more severe depression.

AIMS: This was a post hoc analysis to determine whether baseline severity of depression influenced the efficacy of duloxetine in treating major depressive disorder (MDD) and to better characterise the symptom response profile for duloxetine in patients with mild, moderate or more severe depression. METHODS: Data were pooled from four double-blind, placebo-controlled studies in which outpatients with MDD were randomised to duloxetine (60 mg/day) or placebo for 8-9 weeks. Patients were retrospectively stratified according to baseline 17-item Hamilton Depression Rating scale (HAMD17) total scores: mild=total score

Duloxetine in the treatment of major depressive disorder: a placebo- and paroxetine-controlled trial.

OBJECTIVE: Duloxetine doses of 80 and 120 mg/day were assessed for efficacy and safety in the treatment of major depressive disorder (MDD). METHODS: In this randomized, double-blind trial, patients age > or =18 meeting DSM-IV criteria for MDD were randomized to placebo (N=99), duloxetine 80 mg/day (N=93), duloxetine 120 mg/day (N=103), or paroxetine 20 mg/day (N=97). The primary outcome measure was mean change from baseline in the 17-item Hamilton rating scale for depression (HAMD(17)) total score after 8 weeks of treatment; a number of secondary efficacy measures also were assessed. Safety and tolerability were assessed via collection and analysis of treatment-emergent adverse events (TEAEs), vital signs, and weight. The Arizona sexual experiences scale was used to assess sexual functioning. Patients who had a > or =30% reduction from baseline in the HAMD(17) total score at the end of the acute phase entered a 6-month continuation phase where they remained on the same treatment as they had taken during the acute phase; efficacy and safety/tolerability outcomes were assessed during continuation treatment. RESULTS: More than 87% of patients completed the acute phase in each treatment group. Duloxetine-treated patients (both doses) showed significantly greater improvement (P<0.05) in the HAMD(17) total score at week 8 compared with placebo. Paroxetine was not significantly different from placebo (P=0.089) on mean change on the HAMD(17). Duloxetine 120 mg/day also showed significant improvement on most secondary efficacy measures (six of nine) compared with placebo while duloxetine 80 mg/day (three of nine) and paroxetine (three of nine) were significantly superior to placebo on fewer secondary measures. HAMD(17) mean change data from this study and an identical sister study were pooled as defined a priori for the purposes of performing a non-inferiority test versus paroxetine. Both duloxetine doses met statistical criteria for non-inferiority to paroxetine. TEAE reporting rates were low in all treatment groups and no deaths occurred in the acute or continuation phases. CONCLUSIONS: The efficacy of duloxetine at doses of 80 and 120 mg/day in the treatment of MDD was demonstrated. Tolerability, as measured by TEAEs, and safety were similar to paroxetine 20 mg/day and consistent with previous published data on duloxetine in the treatment of MDD.

Accounting for dropout bias using mixed-effects models.

Treatment effects are often evaluated by comparing change over time in outcome measures. However, valid analyses of longitudinal data can be problematic when subjects discontinue (dropout) prior to completing the study. This study assessed the merits of likelihood-based repeated measures analyses (MMRM) compared with fixed-effects analysis of variance where missing values were imputed using the last observation carried forward approach (LOCF) in accounting for dropout bias. Comparisons were made in simulated data and in data from a randomized clinical trial. Subject dropout was introduced in the simulated data to generate ignorable and nonignorable missingness. Estimates of treatment group differences in mean change from baseline to endpoint from MMRM were, on average, markedly closer to the true value than estimates from LOCF in every scenario simulated. Standard errors and confidence intervals from MMRM accurately reflected the uncertainty of the estimates, whereas standard errors and confidence intervals from LOCF underestimated uncertainty.

Effect of fish oil, arginine, and doxorubicin chemotherapy on remission and survival time for dogs with lymphoma: a double-blind, randomized placebo-controlled study.

BACKGROUND: Polyunsaturated n-3 fatty acids have been shown to inhibit the growth and metastasis of tumors. This double-blind, randomized study was designed to evaluate the hypothesis that polyunsaturated n-3 fatty acids can improve metabolic parameters, decrease chemical indices of inflammation, enhance quality of life, and extend disease free interval and survival time for dogs treated for lymphoblastic lymphoma with doxorubicin chemotherapy. METHODS: Thirty-two dogs with lymphoma were randomized to receive one of two diets supplemented with menhaden fish oil and arginine (experimental diet) or an otherwise identical diet supplemented with soybean oil (control diet). Diets were fed before and after remission was attained with up to five dosages of doxorubicin. Parameters examined included blood concentrations of glucose, lactic acid, and insulin in response to glucose and diet tolerance tests; alpha-1 acid glycoprotein; tumor necrosis factor; interleukin-6; body weight; amino acid profiles; resting energy expenditure; disease free interval (DFI); survival time (ST); and clinical performance scores. RESULTS: Dogs fed the experimental diet had significantly (P < 0.05) higher mean serum levels of the n-3 fatty acids docosahexaenoic acid (C22:6) and eicosapentaenoic acid (C20:5) compared with controls. Higher serum levels of C22:6 and C20:5 were associated with lesser (P < 0.05) plasma lactic acid responses to intravenous glucose and diet tolerance testing. Increasing C22:6 levels were significantly (P < 0.05) associated with longer DFI and ST for dogs with Stage III lymphoma fed the experimental diet. CONCLUSIONS: Fatty acids of the n-3 series normalize elevated blood lactic acid in a dose-dependent manner, resulting in an increase in DFI and ST for dogs with lymphoma.

A bootstrap approach to confidence regions for genetic parameters from Method R estimates.

Confidence regions (CR) for heritability (h2) and fraction of variance accounted for by permanent environmental effects (c2) from Method R estimates were obtained from simulated data using a univariate, repeated measures, full animal model, with 50% subsampling. Bootstrapping techniques were explored to assess the optimum number of subsamples needed to compute Method R estimates of h2 and c2 with properties similar to those of exact estimators. One thousand estimates of each parameter set were used to obtain 90, 95, and 99% CR in four data sets including 2,500 animals with four measurements each. Two approaches were explored to assess CR accuracy: a parametric approach assuming bivariate normality of h2 and c2 and a nonparametric approach based on the sum of squared rank deviations. Accuracy of CR was assessed by the average loss of confidence (LOSS) by number of estimates sampled (NUMEST). For NUMEST = 5, bootstrap estimates of h2 and c2 were within 10(-3) of the asymptotic ones. The same degree of convergence in the estimates of SE was achieved with NUMEST = 20. Correlation between estimates of h2 and c2 ranged from -.83 to -.98. At NUMEST < 10, the nonparametric CR were more accurate than parametric CR. However, with the parametric CR, LOSS approached zero at rate NUMEST(-1). This rate was an order of magnitude larger for the nonparametric CR. These results suggested that when the computational burden of estimating genetic parameters limits the number of Method R estimates that can be obtained to, say, 10 or 20, reliable CR can still be obtained by processing Method R estimates through bootstrapping techniques.

Radiofrequency (electrosurgical) ablation of articular cartilage: a study in sheep.

The objective of this study was to examine the effect of a bipolar ablation probe on experimentally roughened articular cartilage and compare it with the traditional mechanical shaving technique using the knee joint of sheep. Twenty-eight skeletally mature ewes were divided randomly into two groups: one group was treated with a rotating shaving device and another group was treated using the bipolar ablation probe (Bipolar Arthroscopic Probe; Electroscope, Inc, Boulder, CO). Animals were killed at 0, 6, 12, and 24 weeks, and histological sections of the experimental limbs were compared with sections of the opposite limb using a modified Mankin scale. The following variables were used to determine scores: surface (0-6), cells (0-4), hypocellularity (0-3), matrix staining (transitional zone [0-4], radiate zone [0-4], and focal empty lacunae or hypereosinophilic cells (0-3). Differences in scores for all response variables were calculated as treated limb minus sham limb. Response variables were formed: score >0 recoded as 1 (favorable response treated better than sham), score of 0 recoded as 2 (neutral response no differences), and score <0 recoded as 3 (unfavorable response treated worse than sham). Bipolar ablative probe-treated limbs had 14.29% favorable responses and 35.71% favorable or neutral responses, whereas shave-treated limbs had 0% favorable and only 7.14% favorable or neutral responses. For all variables, bipolar ablative probe-treated limbs had more favorable responses. The less severe histological change in the bipolar ablative probe-treated joints compared with the shave-treated joints suggests that bipolar ablation of articular cartilage may be a better treatment for chondromalacia than the usual shaving methods of debridement. Further, there were no pathological changes in the subchondral bone.

Mortality in beagles irradiated during prenatal and postnatal development. II. Contribution of benign and malignant neoplasia.

To evaluate the lifetime carcinogenic hazards of exposure to ionizing radiation during development, 1,680 beagles received whole-body exposures to 60Co gamma rays or sham exposures. Eight groups of 120 dogs each received mean doses of 15.6-17.5 or 80.8-88.3 cGy in early, mid- or late gestation, at 8, 28 or 55 days postcoitus or at 2 days after birth. Another group of 120 dogs received a mean dose of 82.6 cGy as 70-day-old juveniles and one group of 240 dogs received a mean dose of 81.2 cGy as 365-day-old young adults. Sham irradiations were given to 360 controls. Sexes were equally represented. In 1,343 dogs allowed to live out their life span, neoplasia was a major disease, contributing to mortality in 40% of the dogs. There was a significant increase in benign and malignant neoplasms occurring in young dogs (<4 years old), including fatal malignancies, after irradiation in the perinatal (late fetal and neonatal) periods. The lifetime incidence of fatal neoplasms was also increased in dogs irradiated perinatally. Three malignancies-lymphomas, hemangiosarcomas and mammary carcinomas-accounted for 51% of all fatal tumors. There was an apparent lifetime increase and earlier onset of lymphomas in dogs exposed as fetuses. Fatal hemangiosarcomas were increased in dogs irradiated early and late in gestation. Fatal mammary carcinomas were not increased by irradiation, although non-fatal carcinomas were increased after perinatal exposure. Myeloproliferative disorders and central nervous system astrocytomas appeared to be increased in perinatally irradiated dogs. These data suggest that irradiation in both the fetal and neonatal periods is associated with increased early onset and lifetime cancer risk.

Mortality in beagles irradiated during prenatal and postnatal development. I. Contribution of non-neoplastic diseases.

To evaluate the lifetime health effects of exposure to ionizing radiation during development, 1,680 beagles received whole-body exposures to 60Co gamma rays or sham exposures. Eight groups of 120 dogs each received mean doses of 15.6-17.5 or 80.8-88.3 cGy in early, mid- or late gestation, at 8, 28 or 55 days after breeding, or at 2 days after birth. Another group of 120 dogs received a mean dose of 82.6 cGy as 70-day-old juveniles and one group of 240 dogs received a mean dose of 81.2 cGy as 365-day-old young adults. Sham irradiations were given to 360 controls. Sexes were equally represented. There was no significant effect of irradiation on mean survival times in any groups. In 1,343 dogs allowed to live out their life span, chronic renal disease was a common cause of mortality, and irradiation in the late fetal or juvenile periods potentiated this disease, resulting in increased mortality due to renal failure. This was consistent with earlier findings of the high radiosensitivity of the kidney in the perinatal period. Hypothyroidism associated with atrophic thyroiditis was decreased by irradiation, a finding contrary to expectation and not easily explained. Diabetes mellitus was increased by irradiation in the mid- and late gestation and juvenile periods, a finding which is intriguing based on early reports of a similar finding in atomic bomb survivors. Though convulsive seizures were a common cause of mortality in the dogs, there was no evidence for increased risk associated with prenatal irradiation as has been reported in humans. Genetic analyses indicated that renal disease, hypothyroidism, diabetes mellitus and convulsive seizures all had a heritable component, but that this did not influence or bias the radiation responses evaluated.

Duration of effects of dietary fish oil supplementation on serum eicosapentaenoic acid and docosahexaenoic acid concentrations in dogs.

OBJECTIVE: To determine how long serum concentrations of omega-3 fatty acids remain elevated after cessation of dietary fish oil supplementation. ANIMALS: 12 healthy Beagles. PROCEDURE: Baseline serum concentrations of linoleic acid, linolenic acid, arachidonic acid (AA), eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA) were measured. Dogs were then fed a diet supplemented with soybean oil or fish oil for 8 weeks, and serum fatty acid concentrations were measured while dogs were fed the experimental diets and for 18 weeks after they were switched to a maintenance diet. RESULTS: For dogs fed the fish oil diet, serum EPA and DHA concentrations were significantly increased by week 1 and remained increased for 7 (DHA concentration) or 3 (EPA concentration) weeks after dietary fish oil supplementation was discontinued. CONCLUSIONS: In dogs, supplementation of the diet with fish oil may have effects for several weeks after dietary supplementation is discontinued. CLINICAL RELEVANCE: Studies of the effects of fish oil supplementation that use a crossover design should allow for an appropriate washout period.

Approximate confidence intervals for heritability from method R estimates.

Method R estimates of heritability (h2) and associated confidence intervals (CI) were obtained from simulated data using a single trait, direct effects, full animal model, with 50% subsampling. Five hundred data sets were simulated for each of five levels of h2 (.10, .20, .30, .40, and .50) and two types of pedigree structure (random pedigree structure [N = 2,000] that varied over simulations, or the pedigree structure from a real data set [N = 2,644] that was constant for all simulations). The first 10, 20, and all 50 h2 estimates were used to obtain 80, 90, 95, and 99% CI for each data set. The variance of h2 estimates within data sets approximated the sampling variance of the h2 estimates. The Box-Cox transformation was used to normalize the distribution of estimates from each data set. Confidence intervals were computed on the transformed scale as CI = mu +/- (T x sigma), where mu and sigma = the mean and SD of the N transformed h2 estimates, respectively, and T = the critical value from the T distribution for a 1-alpha CI, with df = N-1. Upper and lower CI bounds were converted back to the original scale by reversing the transformation. The percentages of CI containing the true h2 value, pooled across all levels of h2, types of pedigree, and number of estimates used to obtain CI, for 80, 90, 95, and 99% CI were 81.14, 90.96, 95.27, and 98.76%, respectively. These results suggested that Method R h2 estimates can be used to obtain reliable CI.

Non-neoplastic and neoplastic thyroid disease in beagles irradiated during prenatal and postnatal development.

To evaluate the lifetime hazards of exposure to ionizing radiation, 1,680 beagles received whole-body exposures to 60Co gamma rays or sham exposures during development. Eight groups of 120 dogs each received mean doses of 16-18 or 81-88 cGy at 8, 28 or 55 days of gestation, or at 2 days after birth. One group of 120 dogs received a mean of 83 cGy at 70 days of age and one group of 240 dogs received a mean of 81 cGy at 365 days of age. Sham irradiations were given to 360 controls. Sexes were equally represented. In 1,343 dogs allowed to live out their life span, heritable lymphocytic thyroiditis with hypothyroidism was a major contributor to mortality. Irradiated dogs had a decreased risk for hypothyroidism, a finding that was surprising and not easily explained. Of the 1,343 life-span dogs, those exposed as neonates at 2 days of age or as juveniles at 70 days of age had evidence for an increased risk for thyroid follicular cell neoplasia. Hypothyroid dogs had a significantly increased risk for thyroid neoplasia, including greater risk for carcinomas, but no evidence of a greater sensitivity to radiation-induced tumors. In dogs with normal thyroid function irradiated at 2 or 70 days of age there was increased risk for benign and malignant follicular cell neoplasms, including multiple neoplasms. No difference between sexes was noted. These findings related to age sensitivity in the dog were consistent with the high risk for radiogenic thyroid neoplasia in humans after exposure during early childhood.

Doxorubicin alone or in combination with asparaginase, followed by cyclophosphamide, vincristine, and prednisone for treatment of multicentric lymphoma in dogs: 121 cases (1987-1995).

OBJECTIVE: To determine response rate and remission as well as survival times for dogs with multicentric lymphoma treated first with doxorubicin alone or in combination with asparaginase and then with cyclophosphamide, vincristine sulfate, and prednisone (CVP) and to identify prevalence of toxicoses associated with this protocol and factors associated with prognosis. DESIGN: Retrospective case series. ANIMALS: 121 dogs. PROCEDURE: Variables evaluated for prognostic value were initial response rate to chemotherapy, age, breed, sex, body weight, histologic grade, clinical stage and substage, previous corticosteroid treatment, and serum calcium concentration. RESULTS: Median overall remission and survival times for all 121 dogs were 205 and 237 days, respectively. Response rate (complete or partial response) was 88%. Ten dogs were hospitalized because of toxicoses associated with doxorubicin, and 19 dogs were hospitalized because of toxicoses associated with CVP. Asparaginase favorably influenced the initial response rate, but did not significantly influence overall remission of survival times. Initial response rate to chemotherapy, body weight, clinical substage, and serum calcium concentration was found to have prognostic value. CLINICAL IMPLICATIONS: For dogs with multicentric lymphoma, treatment with doxorubicin alone or in combination with asparaginase and then with CVP resulted in an acceptable response rate and low prevalence of toxicoses.

Surgery and doxorubicin in dogs with hemangiosarcoma.

Forty-six dogs with histologically confirmed hemangiosarcoma of various locations other than skin were used in a prospective study to determine the efficacy of adjuvant doxorubicin (30 mg/m2 IV q 3 weeks for 5 treatments) 10 to 14 days after the tumor was partially or completely excised. Analysis of the data included information on variables that were hypothesized to influence response to therapy, disease-free interval (DFI), or survival time (ST). Other information collected included age, gender, breed, weight, prior therapy, type of surgery, location of the primary tumor, presence of metastases, number of doses of doxorubicin, response to doxorubicin therapy (complete or partial response), and the following histological criteria: overall differentiation, nuclear pleomorphism, percent necrosis, mitotic score, total histological score, and grade. Surgery outcome (complete versus incomplete surgical excision) markedly influenced survival times (P < .001). Twenty percent of the dogs rendered free of disease were alive at 1 year, whereas none of the dogs that had residual tumor after surgery were alive at 1 year. Most of the histological criteria (nuclear pleomorphism, mitotic score, grade, overall differentiation) had marked (P < .05), or close to marked, independent associations with ST for dogs that had complete tumor removal. Results from analysis of DFI were generally similar to those of ST in dogs with complete excision of the tumor. Twenty-seven of the 46 dogs (58.7%) had all clinical evidence of tumour successfully removed. Logistic regression analysis of surgical outcome (ability to remove all visible tumor) suggested that age of the subject was the only factor markedly influencing surgical outcome (P = .017). As age increased, the probability of success increased. Those dogs that had previous treatment for their hemangiosarcoma tended (P = .08) to have a shorter DFI and ST. Therefore, complete removal of all evidence of tumor followed by 5 doses of doxorubicin may be an effective treatment for dogs with hemangiosarcoma. Dogs that had all tumor successfully removed had a mean and median ST of 267 and 172 days, respectively. Dogs with incomplete tumor removal had a mean and median ST of 172 and 60 days, respectively. Similarly, prognostic variables such as the ability to completely excise all evidence of tumor, histological criteria, and age of the patient are potentially important prognostic variables for predicting outcome.