About CTProVision.

CT ProVision description.

Subarachnoid Hemorrhage (SAH) consists of spill of blood in the space between the meningeal structures that surround the brain. This phenomenon usually happens after spontaneous rupture of a aneurysmal cerebral artery. Its incidence is estimated at between 4-28/100.000 inhabitants and is the most common cause of sudden death by stroke. Despite the realization of important scientific efforts to improve outcomes for patients who suffer a SAH, the rate of fatal results remains high. In short, it is a serious pathology, with an approximate mortality from 20 to 40% of the patients admitted to the hospital, plus 8 to 15% of mortality in the first few minutes or hours, at the pre-hospital stage.

Most common symptoms and signs usually appear suddenly in an individual who did not usually have previous neurological alterations. The headache, which is the most common symptom, is described as brusque, intense and unusual home. It can also be referred to as nape ache and photophobia. It is often accompanied by nausea and vomiting. But it is not always feature, given that can have any location, can be localized or generalized, can be mild and resolve spontaneously, or be relieved with non-Narcotic analgesics.

Currently, the diagnosis of subarachnoid hemorrhage is performed on the basis of a clinical picture compatible and the presence of blood in the subarachnoid space. The presence of blood in the subarachnoid space is detected by means of an axial tomography, computerized at head level (CT). Precision in the identification of the SAH is very dependent on the quality of computed tomography and the ability of the medical professional who interprets it. It is easy that inexperienced doctors overlooked subtle abnormalities.

The evaluation steps.

• A clinical assessment of the patient according to the Hunt-Hess and WFNS scales, and determine its sex and age.
• Get a picture of computed tomography (CT) of the patient's skull.
• Normalize the image of the skull of the patient according to a template of standard skull.
• Get a region corresponding to intracranial blood through segmentation of the standard image of the patient's brain.
• Calculate at least following parameters of blood intra-craneal region: total volume, its fractal dimension and surface/volume ratio.
• Determine the likelihood of death or vasospasm of the patient on the basis of at least the calculated parameters.

Calculation of probability of death or vasospasm.

For calculation of intra-hospital death, we have built a binary logistic regression model using surface-to-volume ratio of delineated volume of interest, patient´s sex, age and WFNS rating at admission. For calculation of vasospasm we have built a binary logistic regression model using volume of blood, its fractal dimension and Hunt-Hess rating at admission. Clinical data are combined with CT image-derived data to get probability.  

 

Known issues and limitations

Issues:

• If at the end of your calculation passes midnight in GMT+1 time yone (Madrid, Berlin), you may get an error. Try to repeat the calculation.
• If you upload a very large DICOM file (i.e. more than 200 MB) your web session may time out before normalization is finished. Your image is saved, but data cannot be calculated. Contact the developer with patient data to calculate off-line.
• If you use example file and there is a result from somebody´s excercise done the same day on this example file within last 10 min, you get result immediately.
• If you load the same patient twice you get error in calculation and mixed dates in results.
• You need to wait untill calculation is finished in the background to get values and have them saved.

Limitations:

• Calculation of one patient takes usually 2-5 min. For large images of high resolution it may take considerably longer time.
• There is no patient data reading from DICOM file at the moment.
• Patient data saving is starting to develop at the moment. Having more than one study per patient may result in displaying the data of the wrong one or duplicated patients.
• Second study of the same patient (same ID) give unpredictable results or page crash.
• If you miss any data or do not check radiobutton from first, obligatory section, page crashes.

Development path:

• Eliminate the necesity of pressing Calculate Probabilities button, so that the result appears when calculation is finished.
• Patient clinical data saving together with image data. Done in version 0.50
• Each user can review his patients. Done in version 0.50
• Reading patient's data from DICOM file
• Allow editing basic patient data, delete patient
• Allow having more than one study per patient. Done in version 0.50

Version history:

• 0.51 Cross-checking of entered patient data. Minor bug fixes.
• 0.6 Major code revriting and adding of modified Rankin Scale (mRS) to the patient´s data.

Examples

Press the button to pass to Examples page.

---