Is the prefrontal cortex organized by supramodal or modality-specific sensory demands during adolescence?

Attention is inherently biased towards the visual modality during most multisensory scenarios in adults, but the developmental trajectory towards visual dominance has not been fully elucidated. More recent evidence in primates and adult humans suggests a modality-specific stratification of the prefrontal cortex. The current study therefore used functional magnetic resonance imaging (fMRI) to investigate the neuronal correlates of proactive (following cues) and reactive (following probes) cognitive control for simultaneous audio-visual stimulation in 67 healthy adolescents (13-18 years old). Behavioral results were only partially supportive of visual dominance in adolescents, with both reduced response times and accuracy during attend-visual relative to attend-auditory trials. Differential activation of medial and lateral prefrontal cortex for processing incongruent relative to congruent stimuli (reactive control) was also only observed during attend-visual trials. There was no evidence of modality-specific prefrontal cortex stratification during the active processing of multisensory stimuli or during separate functional connectivity analyses. Attention-related modulations were also greater within visual relative to auditory cortex, but were less robust than observed in previous adult studies. Collectively, current results suggest a continued transition towards visual dominance in adolescence, as well as limited modality-specific specialization of prefrontal cortex and attentional modulations of unisensory cortex.

Multimodal Neuroimaging in Schizophrenia: Description and Dissemination.

In this paper we describe an open-access collection of multimodal neuroimaging data in schizophrenia for release to the community. Data were acquired from approximately 100 patients with schizophrenia and 100 age-matched controls during rest as well as several task activation paradigms targeting a hierarchy of cognitive constructs. Neuroimaging data include structural MRI, functional MRI, diffusion MRI, MR spectroscopic imaging, and magnetoencephalography. For three of the hypothesis-driven projects, task activation paradigms were acquired on subsets of ~200 volunteers which examined a range of sensory and cognitive processes (e.g., auditory sensory gating, auditory/visual multisensory integration, visual transverse patterning). Neuropsychological data were also acquired and genetic material via saliva samples were collected from most of the participants and have been typed for both genome-wide polymorphism data as well as genome-wide methylation data. Some results are also presented from the individual studies as well as from our data-driven multimodal analyses (e.g., multimodal examinations of network structure and network dynamics and multitask fMRI data analysis across projects). All data will be released through the Mind Research Network's collaborative informatics and neuroimaging suite (COINS).

Increased spatial granularity of left brain activation and unique age/gender signatures: a 4D frequency domain approach to cerebral lateralization at rest.

Cerebral lateralization is a well-studied topic. However, most of the research to date in functional magnetic resonance imaging (fMRI) has been carried out on hemodynamic fluctuations of voxels, networks, or regions of interest (ROIs). For example, cerebral differences can be revealed by comparing the temporal activation of an ROI in one hemisphere with the corresponding homotopic region in the other hemisphere. While this approach can reveal significant information about cerebral organization, it does not provide information about the full spatiotemporal organization of the hemispheres. The cerebral differences revealed in literature suggest that hemispheres have different spatiotemporal organization in the resting state. In this study, we evaluate cerebral lateralization in the 4D spatiotemporal frequency domain to compare the hemispheres in the context of general activation patterns at different spatial and temporal scales. We use a gender-balanced resting fMRI dataset comprising over 600 healthy subjects ranging in age from 12 to 71, that have previously been studied with a network specific voxel-wise and global analysis of lateralization (Agcaoglu, et al. NeuroImage, 2014). Our analysis elucidates significant differences in the spatiotemporal organization of brain activity between hemispheres, and generally more spatiotemporal fluctuation in the left hemisphere especially in the high spatial frequency bands, and more power in the right hemisphere in the low and middle spatial frequencies. Importantly, the identified effects are not visible in the context of a typical assessment of voxelwise, regional, or even global laterality, thus our study highlights the value of 4D spatiotemporal frequency domain analyses as a complementary and powerful tool for studying brain function.

Lateralization of resting state networks and relationship to age and gender.

Brain lateralization is a widely studied topic, however there has been little work focused on lateralization of intrinsic networks (regions showing similar patterns of covariation among voxels) in the resting brain. In this study, we evaluate resting state network lateralization in an age and gender-balanced functional magnetic resonance imaging (fMRI) dataset comprising over 600 healthy subjects ranging in age from 12 to 71. After establishing sample-wide network lateralization properties, we continue with an investigation of age and gender effects on network lateralization. All data was gathered on the same scanner and preprocessed using an automated pipeline (Scott et al., 2011). Networks were extracted via group independent component analysis (gICA) (Calhoun et al., 2001). Twenty-eight resting state networks discussed in previous (Allen et al., 2011) work were re-analyzed with a focus on lateralization. We calculated homotopic voxelwise measures of laterality in addition to a global lateralization measure, called the laterality cofactor, for each network. As expected, many of the intrinsic brain networks were lateralized. For example, the visual network was strongly right lateralized, auditory network and default mode networks were mostly left lateralized. Attentional and frontal networks included nodes that were left lateralized and other nodes that were right lateralized. Age was strongly related to lateralization in multiple regions including sensorimotor network regions precentral gyrus, postcentral gyrus and supramarginal gyrus; and visual network regions lingual gyrus; attentional network regions inferior parietal lobule, superior parietal lobule and middle temporal gyrus; and frontal network regions including the inferior frontal gyrus. Gender showed significant effects mainly in two regions, including visual and frontal networks. For example, the inferior frontal gyrus was more right lateralized in males. Significant effects of age were found in sensorimotor and visual networks on the global measure. In summary, we report a large-sample of lateralization study that finds intrinsic functional brain networks to be highly lateralized, with regions that are strongly related to gender and age locally, and with age a strong factor in lateralization, and gender exhibiting a trend-level effect on global measures of laterality.

Hippocampal structural and functional changes associated with electroconvulsive therapy response.

Previous animal models and structural imaging investigations have linked hippocampal neuroplasticity to electroconvulsive therapy (ECT) response, but the relationship between changes in hippocampal volume and temporal coherence in the context of ECT response is unknown. We hypothesized that ECT response would increase both hippocampal resting-state functional magnetic resonance imaging connectivity and hippocampal volumes. Patients with major depressive disorder (n=19) were scanned before and after the ECT series. Healthy, demographically matched comparisons (n=20) were scanned at one-time interval. Longitudinal changes in functional connectivity of hippocampal regions and volumes of hippocampal subfields were compared with reductions in ratings of depressive symptoms. Right hippocampal connectivity increased (normalized) after the ECT series and correlated with depressive symptom reduction. Similarly, the volumes of the right hippocampal cornu ammonis (CA2/3), dentate gyrus and subiculum regions increased, but the hippocampal subfields were unchanged relative to the comparison group. Connectivity changes were not evident in the left hippocampus, and volume changes were limited to the left CA2/3 subfields. The laterality of the right hippocampal functional connectivity and volume increases may be related to stimulus delivery method, which was predominately right unilateral in this investigation. The findings suggested that increased hippocampal functional connectivity and volumes may be biomarkers for ECT response.

A practical approach to incidental findings in neuroimaging research.

OBJECTIVE: We describe the systematic approach to incidental findings (IFs) used at the Mind Research Network (MRN) where all MRI scans receive neuroradiologist interpretation and participants are provided results. METHODS: From 2004 to 2011, 8,545 MRI scans were acquired by 45 researchers. As mandated by MRN's external institutional review board, all structural sequences were evaluated by a clinical neuroradiologist who generated a report that included recommendations for referral if indicated. Investigators received a copy of their participants' reports, which were also mailed to participants unless they specifically declined. To better understand the impact of the radiology review process, a financial analysis was completed in addition to a follow-up phone survey to characterize participant perceptions regarding receiving their MRI scan results. RESULTS: The radiologist identified IFs in 34% of the 4,447 participants. Of those with IFs (n = 1,518), the radiologist recommended urgent or immediate referral for 2.5% and routine referral for 17%. For 80.5%, no referral was recommended. Estimated annual cost for this approach including support for the neuroradiologist, medical director, and ancillary staff is approximately $60,000 or $24/scan. The results of the retrospective phone survey showed that 92% of participants appreciated receiving their MRI report, and the majority stated it increased their likelihood of volunteering for future studies. CONCLUSIONS: Addressing IFs in a cost-effective and consistent manner is possible by adopting a policy that provides neuroradiology interpretation and offers participant assistance with clinical follow-up when necessary. Our experience suggests that an ethical, institution-wide approach to IFs can be implemented with minimal investigator burden.

A prospective diffusion tensor imaging study in mild traumatic brain injury.

OBJECTIVES: Only a handful of studies have investigated the nature, functional significance, and course of white matter abnormalities associated with mild traumatic brain injury (mTBI) during the semi-acute stage of injury. The present study used diffusion tensor imaging (DTI) to investigate white matter integrity and compared the accuracy of traditional anatomic scans, neuropsychological testing, and DTI for objectively classifying mTBI patients from controls. METHODS: Twenty-two patients with semi-acute mTBI (mean = 12 days postinjury), 21 matched healthy controls, and a larger sample (n = 32) of healthy controls were studied with an extensive imaging and clinical battery. A subset of participants was examined longitudinally 3-5 months after their initial visit. RESULTS: mTBI patients did not differ from controls on clinical imaging scans or neuropsychological performance, although effect sizes were consistent with literature values. In contrast, mTBI patients demonstrated significantly greater fractional anisotropy as a result of reduced radial diffusivity in the corpus callosum and several left hemisphere tracts. DTI measures were more accurate than traditional clinical measures in classifying patients from controls. Longitudinal data provided preliminary evidence of partial normalization of DTI values in several white matter tracts. CONCLUSIONS: Current findings of white matter abnormalities suggest that cytotoxic edema may be present during the semi-acute phase of mild traumatic brain injury (mTBI). Initial mechanical damage to axons disrupts ionic homeostasis and the ratio of intracellular and extracellular water, primarily affecting diffusion perpendicular to axons. Diffusion tensor imaging measurement may have utility for objectively classifying mTBI, and may serve as a potential biomarker of recovery.

The effects of stimulus modality and frequency of stimulus presentation on cross-modal distraction.

Selective attention produces enhanced activity (attention-related modulations [ARMs]) in cortical regions corresponding to the attended modality and suppressed activity in cortical regions corresponding to the ignored modality. However, effects of behavioral context (e.g., temporal vs. spatial tasks) and basic stimulus properties (i.e., stimulus frequency) on ARMs are not fully understood. The current study used functional magnetic resonance imaging to investigate selectively attending and responding to either a visual or auditory metronome in the presence of asynchronous cross-modal distractors of 3 different frequencies (0.5, 1, and 2 Hz). Attending to auditory information while ignoring visual distractors was generally more efficient (i.e., required coordination of a smaller network) and less effortful (i.e., decreased interference and presence of ARMs) than attending to visual information while ignoring auditory distractors. However, these effects were modulated by stimulus frequency, as attempting to ignore auditory information resulted in the obligatory recruitment of auditory cortical areas during infrequent (0.5 Hz) stimulation. Robust ARMs were observed in both visual and auditory cortical areas at higher frequencies (2 Hz), indicating that participants effectively allocated attention to more rapidly presented targets. In summary, results provide neuroanatomical correlates for the dominance of the auditory modality in behavioral contexts that are highly dependent on temporal processing.

The neural networks underlying auditory sensory gating.

One of the most consistent electrophysiological deficits reported in the schizophrenia literature is the failure to inhibit, or properly gate, the neuronal response to the second stimulus of an identical pair (i.e., sensory gating). Although animal and invasive human studies have consistently implicated the auditory cortex, prefrontal cortex and hippocampus in mediating the sensory gating response, localized activation in these structures has not always been reported during non-invasive imaging modalities. In the current experiment, event-related FMRI and a variant of the traditional gating paradigm were utilized to examine how the gating network differentially responded to the processing of pairs of identical and non-identical tones. Two single-tone conditions were also presented so that they could be used to estimate the HRF for paired stimuli, reconstructed based on actual hemodynamic responses, to serve as a control non-gating condition. Results supported an emerging theory that the gating response for both paired-tone conditions was primarily mediated by auditory and prefrontal cortex, with potential contributions from the thalamus. Results also indicated that the left auditory cortex may play a preferential role in determining the stimuli that should be inhibited (gated) or receive further processing due to novelty of information. In contrast, there was no evidence of hippocampal involvement, suggesting that future work is needed to determine what role it may play in the gating response.

Somatotopic organization of the medial wall of the cerebral hemispheres: a 3 Tesla fMRI study.

Somatotopic organization of the human medial wall of the cerebral hemispheres was studied using functional MRI conducted at high field strength (3 T) with fine spatial resolution ( approximately 2 mm). Healthy subjects performed paced, repetitive movements of the fingers and toes. Within the supplementary motor area (SMA), two regions were identified: finger movements activated a region rostral and superior to that for toe movements. Two activation foci were also identified in the cingulate motor area: toe movements activated a region rostral and ventral to that activated by finger movements. All foci were located between the anterior and posterior commissures. Our results confirm previous human and non-human primate studies regarding the rostral-caudal organization of the SMA and CMA. The dorsal-ventral organization of the CMA, however, appears to be divergent from results derived from cortical stimulation studies conducted in non-human primates.

Can medial temporal lobe regions distinguish true from false? An event-related functional MRI study of veridical and illusory recognition memory.

To investigate the types of memory traces recovered by the medial temporal lobe (MTL), neural activity during veridical and illusory recognition was measured with the use of functional MRI (fMRI). Twelve healthy young adults watched a videotape segment in which two speakers alternatively presented lists of associated words, and then the subjects performed a recognition test including words presented in the study lists (True items), new words closely related to studied words (False items), and new unrelated words (New items). The main finding was a dissociation between two MTL regions: whereas the hippocampus was similarly activated for True and False items, suggesting the recovery of semantic information, the parahippocampal gyrus was more activated for True than for False items, suggesting the recovery of perceptual information. The study also yielded a dissociation between two prefrontal cortex (PFC) regions: whereas bilateral dorsolateral PFC was more activated for True and False items than for New items, possibly reflecting monitoring of retrieved information, left ventrolateral PFC was more activated for New than for True and False items, possibly reflecting semantic processing. Precuneus and lateral parietal regions were more activated for True and False than for New items. Orbitofrontal cortex and cerebellar regions were more activated for False than for True items. In conclusion, the results suggest that activity in anterior MTL regions does not distinguish True from False, whereas activity in posterior MTL regions does.

Paclitaxel, an active agent in nonsquamous carcinomas of the uterine cervix: a Gynecologic Oncology Group Study.

PURPOSE: A phase II trial of paclitaxel was initiated in advanced nonsquamous carcinoma of the cervix to determine its activity in patients who had failed standard chemotherapy. PATIENTS AND METHODS: Eligible patients had at least one measurable lesion. The starting dose of paclitaxel was 170 mg/m(2) (135 mg/m(2) for patients with prior pelvic radiation) given as a 24-hour continuous intravenous infusion with courses repeated every 3 weeks. Dose escalation to 200 mg/m(2) and de-escalation to 110 mg/m(2) were allowed based on adverse effects. RESULTS: In this trial, 42 assessable patients were initially entered onto the study, and 13 responses were seen; four patients had a complete response, and nine patients had a partial response. The overall response rate was 31%. The primary and dose-limiting toxicity was neutropenia. CONCLUSION: The response rate to paclitaxel exceeds the rates reported using other single agents in nonsquamous carcinoma of the cervix.

The evolution of brain activation during temporal processing.

Timing is crucial to many aspects of human performance. To better understand its neural underpinnings, we used event-related fMRI to examine the time course of activation associated with different components of a time perception task. We distinguished systems associated with encoding time intervals from those related to comparing intervals and implementing a response. Activation in the basal ganglia occurred early, and was uniquely associated with encoding time intervals, whereas cerebellar activation unfolded late, suggesting an involvement in processes other than explicit timing. Early cortical activation associated with encoding of time intervals was observed in the right inferior parietal cortex and bilateral premotor cortex, implicating these systems in attention and temporary maintenance of intervals. Late activation in the right dorsolateral prefrontal cortex emerged during comparison of time intervals. Our results illustrate a dynamic network of cortical-subcortical activation associated with different components of temporal information processing.

Specialized neural systems underlying representations of sequential movements.

The ease by which movements are combined into skilled actions depends on many factors, including the complexity of movement sequences. Complexity can be defined by the surface structure of a sequence, including motoric properties such as the types of effectors, and by the abstract or sequence-specific structure, which is apparent in the relations amongst movements, such as repetitions. It is not known whether different neural systems support the cognitive and the sensorimotor processes underlying different structural properties of sequential actions. We investigated this question using whole-brain functional magnetic resonance imaging (fMRI) in healthy adults as they performed sequences of five key presses involving up to three fingers. The structure of sequences was defined by two factors that independently lengthen the time to plan sequences before movement: the number of different fingers (1-3; surface structure) and the number of finger transitions (0-4; sequence-specific structure). The results showed that systems involved in visual processing (extrastriate cortex) and the preparation of sensory aspects of movement (rostral inferior parietal and ventral premotor cortex (PMv)) correlated with both properties of sequence structure. The number of different fingers positively correlated with activation intensity in the cerebellum and superior parietal cortex (anterior), systems associated with sensorimotor, and kinematic representations of movement, respectively. The number of finger transitions correlated with activation in systems previously associated with sequence-specific processing, including the inferior parietal and the dorsal premotor cortex (PMd), and in interconnecting superior temporal-middle frontal gyrus networks. Different patterns of activation in the left and right inferior parietal cortex were associated with different sequences, consistent with the speculation that sequences are encoded using different mnemonics, depending on the sequence-specific structure. In contrast, PMd activation correlated positively with increases in the number of transitions, consistent with the role of this area in the retrieval or preparation of abstract action plans. These findings suggest that the surface and the sequence-specific structure of sequential movements can be distinguished by distinct distributed systems that support their underlying mental operations.

Neural systems underlying the recognition of familiar and newly learned faces.

Memory for famous faces can be used to examine the neural systems underlying retrieval from long-term memory. To date, there have been a limited number of functional neuroimaging investigations examining famous face recognition. In this study, we compared recognition of famous faces to recognition of newly learned faces. Whole-brain, event-related functional magnetic resonance imaging was used to image regional changes in neural activity in 11 subjects during the encoding of unfamiliar faces and during familiarity judgments for: (1) newly learned faces, (2) unfamiliar face distractors, and (3) famous faces. Image analyses were restricted to correct recognition trials. Recognition accuracy and response time to famous and recently learned faces were equivalent. Recognition of famous faces was associated with a widespread network of bilateral brain activations involving the prefrontal, lateral temporal, and mesial temporal (hippocampal and parahippocampal regions) regions compared to recognition of recently encoded faces or unfamiliar faces seen for the first time. Findings are discussed in relation to current proposals concerning the neural regions thought to participate in long-term memory retrieval and, more specifically, in relation to retrieval of information from the person identity semantic system.

Handedness and psychopathy.

OBJECTIVES: The current study investigated whether psychopathic persons exhibit a pattern of motor dominance that differs from the left-hemisphere dominance that characterizes the majority of the population. BACKGROUND: Research with cognitive tasks suggests that psychopathy may be associated with an atypical pattern of cerebral organization characterized by either left-hemisphere dysfunction or decreased hemispheric asymmetry. However, studies that evaluate motor dominance in psychopathy have provided mixed results. METHOD: The handedness of 420 adult male inmates at a county jail was assessed using a scale developed by Chapman and Chapman as part of their participation in ongoing research. RESULTS: Psychopaths reported reduced right-hand dominance, which cannot be accounted for by differences in age, intelligence, or race. CONCLUSIONS: Psychopathy is associated with an increased proportion of mixed-handedness relative to that seen in nonpsychopathic persons. This finding is consistent with theory and research suggesting anomalous cerebral asymmetry in psychopathic offenders.

Neural mechanisms of visual attention: object-based selection of a region in space.

Objects play an important role in guiding spatial attention through a cluttered visual environment. We used event-related functional magnetic resonance imaging (ER-fMRI) to measure brain activity during cued discrimination tasks requiring subjects to orient attention either to a region bounded by an object (object-based spatial attention) or to an unbounded region of space (location-based spatial attention) in anticipation of an upcoming target. Comparison between the two tasks revealed greater activation when attention selected a region bounded by an object. This activation was strongly lateralized to the left hemisphere and formed a widely distributed network including (a) attentional structures in parietal and temporal cortex and thalamus, (b) ventral-stream object processing structures in occipital, inferior-temporal, and parahippocampal cortex, and (c) control structures in medial- and dorsolateral-prefrontal cortex. These results suggest that object-based spatial selection is achieved by imposing additional constraints over and above those processes already operating to achieve selection of an unbounded region. In addition, ER-fMRI methodology allowed a comparison of validly versus invalidly cued trials, thereby delineating brain structures involved in the reorientation of attention after its initial deployment proved incorrect. All areas of activation that differentiated between these two trial types resulted from greater activity during the invalid trials. This outcome suggests that all brain areas involved in attentional orienting and task performance in response to valid cues are also involved on invalid trials. During invalid trials, additional brain regions are recruited when a perceiver recovers from invalid cueing and reorients attention to a target appearing at an uncued location. Activated brain areas specific to attentional reorientation were strongly right-lateralized and included posterior temporal and inferior parietal regions previously implicated in visual attention processes, as well as prefrontal regions that likely subserve control processes, particularly related to inhibition of inappropriate responding.

Chemical analysis of adnexal cyst fluid.

OBJECTIVE: The purpose of the study was to determine if adnexal cyst fluid glucose, protein, and lactate dehydrogenase (LDH) levels can discriminate between cancerous and noncancerous adnexal masses. METHODS: Intracystic fluid was aspirated from adnexal masses immediately after operative excision. Patient age and menopausal status, mass size, and cyst fluid specific gravity were recorded. Cyst fluid and intraoperative serum glucose, protein, and LDH levels were measured. Masses were grouped by histopathologic diagnosis. Cyst fluid chemical levels and cyst fluid/serum ratios were compared among and between the groups. RESULTS: Fifty-eight adnexal masses were analyzed: 15 nonneoplastic (group 1), 23 benign neoplastic (group 2), and 20 malignant (group 3). There were no significant differences among the groups with regard to patient age, menopausal status, or cyst fluid specific gravity. Cyst size (cm2) was significantly different among the three groups (P < 0.01), with the largest mean size found in the cancer group. No significant differences in cyst chemistries or cyst fluid/serum ratios were found between groups 1 and 2. Comparing groups 1 and 3, all values were significantly different (P < 0.05), with the greatest level of significance attained by comparison of cyst fluid LDH levels (P < 0.001). Groups 2 and 3 statistically differed in cyst fluid levels and cyst fluid/serum ratios of both protein and LDH, with the highest levels of significance achieved by comparisons of cyst fluid levels and ratios of LDH (P = 0.001 and P < 0.001, respectively). The cyst fluid LDH level was found to be the best single chemistry for distinguishing noncancerous (groups 1 and 2) from cancerous (group 3) adnexal masses. A cyst fluid LDH level of >/=451 U/L imparted a 90% sensitivity and 71% specificity for detecting malignancy. CONCLUSIONS: Evaluation of adnexal cyst fluid LDH may help to distinguish benign from malignant adnexal masses. More cases are needed to adequately assess the predictive value and clinical utility of this approach.

Prolonged oral etoposide as second-line therapy for platinum-resistant and platinum-sensitive ovarian carcinoma: a Gynecologic Oncology Group study.

PURPOSE: A phase II trial was conducted to determine the activity of prolonged oral etoposide in platinum-resistant and platinum-sensitive ovarian carcinoma. PATIENTS AND METHODS: Platinum-resistant disease was defined as progression on platinum-based chemotherapy or recurrence within 6 months of completing therapy. The starting dose was 50 mg/m2/d (30 mg/m2/d for prior radiotherapy) for 21 days, every 28 days. A dose escalation to a maximum dose of 60 mg/m2/d was prescribed. RESULTS: Of 99 patients entered, 97 were assessable for toxicity and 82 were assessable for response. Among 41 platinum-resistant patients a 26.8% response rate (7.3% complete response [CR] and 19.5% partial response [PR] rate) occurred. The median response duration was 4.3 months (range, 1.3 to 8.7), median progression-free interval (PFI) was 5.7 months (range, 0.8 to 30.8+), and median survival time was 10.8 months (range, 1.9 to 45.8). Twenty-five of 41 platinum-resistant patients had also previously received paclitaxel; of which eight (32%) responded. Among 41 platinum-sensitive patients, a 34.1% response rate (14.6% CR and 19.5% PR rate) occurred. The median response duration was 7.5 months (range, 1.9 to 15.2+), median PFI was 6.3+ months (range, 0.9 to 20.4), and median survival time was 16.5+ months (range, 0.9 to 34.8). Of 97 patients assessable for toxicity, grade 3 or 4 hematologic toxicity was common, with leukopenia occurring in 41.2% (grade 3, 29%; grade 4, 12%), neutropenia in 45.4% (grade 3, 20%; grade 4, 25%), thrombocytopenia in 9% (grade 3, 5%; grade 4, 4%), and anemia in 13.4%. Three treatment-related deaths occurred: two from neutropenic sepsis and one from thrombocytopenic bleeding after an overdose. One patient developed leukemia. CONCLUSION: This regimen is active in platinum-resistant and platinum-sensitive ovarian carcinoma. Additionally, the regimen is active in paclitaxel-resistant ovarian carcinoma.