Unsupervised Deep Learning for Stroke Lesion Segmentation on Follow-up CT Based on Generative Adversarial Networks.

BACKGROUND AND PURPOSE: Supervised deep learning is the state-of-the-art method for stroke lesion segmentation on NCCT. Supervised methods require manual lesion annotations for model development, while unsupervised deep learning methods such as generative adversarial networks do not. The aim of this study was to develop and evaluate a generative adversarial network to segment infarct and hemorrhagic stroke lesions on follow-up NCCT scans. MATERIALS AND METHODS: Training data consisted of 820 patients with baseline and follow-up NCCT from 3 Dutch acute ischemic stroke trials. A generative adversarial network was optimized to transform a follow-up scan with a lesion to a generated baseline scan without a lesion by generating a difference map that was subtracted from the follow-up scan. The generated difference map was used to automatically extract lesion segmentations. Segmentation of primary hemorrhagic lesions, hemorrhagic transformation of ischemic stroke, and 24-hour and 1-week follow-up infarct lesions were evaluated relative to expert annotations with the Dice similarity coefficient, Bland-Altman analysis, and intraclass correlation coefficient. RESULTS: The median Dice similarity coefficient was 0.31 (interquartile range, 0.08-0.59) and 0.59 (interquartile range, 0.29-0.74) for the 24-hour and 1-week infarct lesions, respectively. A much lower Dice similarity coefficient was measured for hemorrhagic transformation (median, 0.02; interquartile range, 0-0.14) and primary hemorrhage lesions (median, 0.08; interquartile range, 0.01-0.35). Predicted lesion volume and the intraclass correlation coefficient were good for the 24-hour (bias, 3 mL; limits of agreement, -64-59 mL; intraclass correlation coefficient, 0.83; 95% CI, 0.78-0.88) and excellent for the 1-week (bias, -4 m; limits of agreement,-66-58 mL; intraclass correlation coefficient, 0.90; 95% CI, 0.83-0.93) follow-up infarct lesions. CONCLUSIONS: An unsupervised generative adversarial network can be used to obtain automated infarct lesion segmentations with a moderate Dice similarity coefficient and good volumetric correspondence.

Prediction of Outcome Using Quantified Blood Volume in Aneurysmal SAH.

BACKGROUND AND PURPOSE: In patients with SAH, the amount of blood is strongly associated with clinical outcome. However, it is commonly estimated with a coarse grading scale, potentially limiting its predictive value. Therefore, we aimed to develop and externally validate prediction models for clinical outcome, including quantified blood volumes, as candidate predictors. MATERIALS AND METHODS: Clinical and radiologic candidate predictors were included in a logistic regression model. Unfavorable outcome was defined as a modified Rankin Scale score of 4-6. An automatic hemorrhage-quantification algorithm calculated the total blood volume. Blood was manually classified as cisternal, intraventricular, or intraparenchymal. The model was selected with bootstrapped backward selection and validated with the R 2, C-statistic, and calibration plots. If total blood volume remained in the final model, its performance was compared with models including location-specific blood volumes or the modified Fisher scale. RESULTS: The total blood volume, neurologic condition, age, aneurysm size, and history of cardiovascular disease remained in the final models after selection. The externally validated predictive accuracy and discriminative power were high (R 2 = 56% ± 1.8%; mean C-statistic = 0.89 ± 0.01). The location-specific volume models showed a similar performance (R 2 = 56% ± 1%, P = .8; mean C-statistic = 0.89 ± 0.00, P = .4). The modified Fisher models were significantly less accurate (R 2 = 45% ± 3%, P < .001; mean C-statistic = 0.85 ± 0.01, P = .03). CONCLUSIONS: The total blood volume-based prediction model for clinical outcome in patients with SAH showed a high predictive accuracy, higher than a prediction model including the commonly used modified Fisher scale.

Association of Quantified Location-Specific Blood Volumes with Delayed Cerebral Ischemia after Aneurysmal Subarachnoid Hemorrhage.

BACKGROUND AND PURPOSE: Delayed cerebral ischemia is a severe complication of aneurysmal SAH and is associated with a high case morbidity and fatality. The total blood volume and the presence of intraventricular blood on CT after aneurysmal SAH are associated with delayed cerebral ischemia. Whether quantified location-specific (cisternal, intraventricular, parenchymal, and subdural) blood volumes are associated with delayed cerebral ischemia has been infrequently researched. This study aimed to associate quantified location-specific blood volumes with delayed cerebral ischemia. MATERIALS AND METHODS: Clinical and radiologic data were collected retrospectively from consecutive patients with aneurysmal SAH with available CT scans within 24 hours after ictus admitted to 2 academic centers between January 2009 and December 2011. Total blood volume was quantified using an automatic hemorrhage-segmentation algorithm. Segmented blood was manually classified as cisternal, intraventricular, intraparenchymal, or subdural. Adjusted ORs with 95% confidence intervals for delayed cerebral ischemia per milliliter of location-specific blood were calculated using multivariable logistic regression analysis. RESULTS: We included 282 patients. Per milliliter increase in blood volume, the adjusted OR for delayed cerebral ischemia was 1.02 (95% CI, 1.01-1.04) for cisternal, 1.02 (95% CI, 1.00-1.04) for intraventricular, 0.99 (95% CI, 0.97-1.02) for intraparenchymal, and 0.96 (95% CI, 0.86-1.07) for subdural blood. CONCLUSIONS: Our findings suggest that in patients with aneurysmal subarachnoid hemorrhage, the cisternal blood volume has a stronger relation with delayed cerebral ischemia than the blood volumes at other locations in the brain.

Ruptured middle cerebral artery aneurysms with a concomitant intraparenchymal hematoma: the role of hematoma volume.

PURPOSE: To study whether clinical outcome data from our patient cohort could give support to the new recommendation in the AHA/ASA guidelines for the management of aneurysmal subarachnoid hemorrhage that states "that microsurgical clipping may receive increased consideration in patients with ruptured middle cerebral artery (MCA) aneurysms and large (>50 mL) intraparenchymal hematomas", while clinical outcome data supporting this recommendation are sparse. METHODS: We reviewed the clinical and radiological data of 81 consecutive patients with MCA aneurysms and concomitant hematomas admitted between January 2006 and December 2015. The relation between (semi-automatically quantified) hematoma volume (< or > 50 ml), neurological condition on admission (poor: GCS < 8 or non-reactive pupils), treatment strategies (no treatment, coiling, or clipping with or without decompression and/or clot removal), and outcome (favorable: mRS score 0-3) was evaluated. RESULTS: Clinical outcome data were available for 76 patients. A significant difference in favorable outcome (17 vs 68%) was seen when comparing patients with poor and good neurological condition on admission (p < 0.01). Patients with hematomas > 50 ml had similar outcomes for coiling and clipping, all underwent decompression. Patients with hematomas < 50 ml did not show differences in favorable outcome when comparing coiling and clipping with (33 and 31%) or without decompression (90 and 88%). CONCLUSION: Poor neurological condition on admission, and not large intraparenchymal hematoma volume, was associated with poor clinical outcome. Therefore, even in patients with large hematomas, the neurological condition on admission and the aneurysm configuration seem to be equally important factors to determine the most appropriate treatment strategy.

Drugs versus diets: disillusions with Dutch health care.

Biology incorporated into other disciplines is often distorted, alarmingly so in some areas of medicine. Together with other forms of bias, this may have detrimental effects for patients depending on medical research for their health. A case study concerning omeprazole (Losec), one of the acid-suppressive drugs against gastric ulcers, and NSAIDs, non-steroid anti-inflammatory drugs, confirms that distorted biology together with biased health care policies foster disasters in current biomedicine and medical practice. In our country, The Netherlands, omeprazole is presumably the most commonly used medication. NSAIDs are also used in large quantities, increasingly since they have become available as analgesic over-the-counter drugs. Unofficial and official sources tend to inform the general public that the drugs promote human health. We argue that their being used on a massive scale is actually a medical disaster. The health of many patients would be served better if the drugs they take were replaced by proper forms of diet, but the pharmaceutical industry, the most potent force affecting medication policies, appears to prevent a shift in the balance from over-medicalization towards healthy life styles. The shift should come from government agencies responsible for regulation in the medication market. Policies of these agencies are now a dismal failure.

Chronic fatigue syndrome: a matter of enzyme deficiencies?

The etiology of chronic fatigue syndrome (CFS) remains an enigma. But literature concerning chronic fatigue which does not focus on CFS points to all sorts of enzyme deficiencies as possible causes. The deficiencies are probably dismissed as causes of CFS because other characteristic symptoms are lacking in CFS patients. But these symptoms are often also lacking in patients with a deficiency. Symptom patterns in enzyme deficiencies are extremely variable. Therefore, patients with CFS should be screened systematically for enzyme deficiencies.

Bias in behaviour genetics: an ecological perspective.

Research in behaviour genetics uncovers causes of behaviour at the population level. For inferences about individuals we also need to know how genes and the environment affect phenotypes. Behaviour genetics fosters a biased view of individual behaviour since it identifies the environment with psychosocial factors and disregards ecology.

New ways to look at fitness.

Many authors have argued that the core of evolutionary biology as represented by the catchphrase 'The fittest survive' is tautological. Concerning the fitness concept of population genetics it is easy to rebut this charge by a proper explication of the term 'survival'. In biology and in the philosophy of biology, various fitness concepts over and above that of population genetics have been elaborated. These concepts, which are called 'supervenient' by some philosophers, have a limited usefulness. On some interpretations they do lead to unacceptable tautologies and circular reasoning. The so-called propensity concept of fitness is problematic in this respect. If interpreted in a proper way, supervenient concepts appear not to allow the formulation of highly general explanations and theories; at best they reveal common patterns among diverse non-general explanations and theories. Philosophers cherishing supervenient fitness concepts are apparently motivated by a mistaken search for general theory. The fitness concept of population genetics may play a role in relatively general theories and explanations. Supervenient concepts cannot play such a role. They should rather help us recognize the value of natural history in biology.

Methodological problems in evolutionary biology. V. The import of supervenience.

Rosenberg has rightly argued that fitness is supervenient. But he has wrongly assumed that this makes "The fittest survive" nontautologous. Supervenience makes strict reduction impossible. It sheds light on disputes concerning the testability of evolutionary theory.

Methodological problems in evolutionary biology. VI. The force of evolutionary epistemology.

Evolutionary epistemology takes various forms. As a philosophical discipline, it may use analogies by borrowing concepts from evolutionary biology to establish new foundations. This is not a very successful enterprise because the analogies involved are so weak that they hardly have explanatory force. It may also veil itself with the garbs of biology. Proponents of this strategy have only produced irrelevant theories by transforming epistemology's concepts beyond recognition. Sensible theories about "knowledge and biology" should presuppose that various long-standing problems concerning relations between the mental and the physical are solved. Such problems are wrongly disregarded by evolutionary epistemologists.

Methodological problems in evolutionary biology. VII. The species plague.

Various philosophers and evolutionary biologists have recently defended the thesis that species are individuals rather than sets. A decade of debates, however, did not suffice to settle the matter. Conceptual analysis shows that many of the key terms involved ("individuation", "evolutionary species", "spatiotemporal restrictedness", "individual") are ambiguous. Current disagreements should dissolve once this is recognized. Explication of the concepts involved leads to new programs for philosophical research. It could also help biology by showing how extant controversies concerning evolution may have conceptual rather than factual roots.