Reflectance Estimation from Multispectral Linescan Acquisitions under Varying Illumination-Application to Outdoor Weed Identification.

To reduce the amount of herbicides used to eradicate weeds and ensure crop yields, precision spraying can effectively detect and locate weeds in the field thanks to imaging systems. Because weeds are visually similar to crops, color information is not sufficient for effectively detecting them. Multispectral cameras provide radiance images with a high spectral resolution, thus the ability to investigate vegetated surfaces in several narrow spectral bands. Spectral reflectance has to be estimated in order to make weed detection robust against illumination variation. However, this is a challenge when the image is assembled from successive frames that are acquired under varying illumination conditions. In this study, we present an original image formation model that considers illumination variation during radiance image acquisition with a linescan camera. From this model, we deduce a new reflectance estimation method that takes illumination at the frame level into account. We experimentally show that our method is more robust against illumination variation than state-of-the-art methods. We also show that the reflectance features based on our method are more discriminant for outdoor weed detection and identification.

Systematic Review and Meta-Analysis of Preoperative Risk Factors of Type II Endoleaks after Endovascular Aneurysm Repair.

BACKGROUND: Type II endoleaks (T2Es) remain the Achilles heel of endovascular aneurysm repair (EVAR), involving a close follow-up and sometimes leading to reintervention. Identifying risk factors impacting T2Es is of concern to improve decision making and optimize follow-up. However, it has led to contradictory results, with supporting evidence for the majority of factors being weak. METHODS: A systematic review and meta-analysis was conducted to study risk factors of T2Es following EVAR to identify risk factors and measure their dedicated strength of association. Using a literature search of MEDLINE, EMBASE, and the Cochrane Library, 31 retrospective studies including a total of 15,793 patients were identified and fulfilled the strict specified inclusion criteria. Random-effects meta-analysis was conducted for each factor to combine effect estimate across studies. A total of 21 factors related to demography, preoperative treatment, comorbidity, and morphology were statistically pooled. RESULTS: On the basis of the pooled odds ratios and their 95% confidence intervals, patency of aortic side branches, represented by the patency of the inferior mesenteric artery, lumbar arteries, or total number of aortic side branches, were found to be significant harmful risk factors of T2Es. Women were also found to have nearly significant higher risk of developing T2Es than men. On the contrary, the following were found to have a significant protective role: smoking, peripheral artery disease, and thrombus load, represented by the maximum thickness at the maximum aneurysm diameter, the presence of circumferential thrombus, or the presence of thrombus at the level of inferior mesenteric artery. CONCLUSION: Identifying significant risk factors of development of T2Es is mandatory to improve decision making and optimize surveillance planning in EVAR.

Predictive Models of Complications after Endovascular Aortic Aneurysm Repair.

BACKGROUND: The risk of long-term complications after endovascular aneurysm repair (EVAR) is still higher than open surgery and is a critical issue. This study aims to make available reliable statistical predictive models of complications after EVAR. METHODS: Two hundred and thirteen patients who underwent EVAR between 2002 and 2012 were included in this study. The preoperative computed tomography scans were analyzed with a dedicated workstation to provide spatially correct 3-dimensional data. Age, gender, operation-related factors, and 21 morphologic variables were measured and included in the analyses. Five postoperative outcomes were studied. After an initial selection of predictors based on univariate analysis, binomial logistic regression models were proposed for each outcome. The ability to predict each outcome was assessed with receiver operating characteristic curves considering that an area under the curve (AUC) > 0.70 is generally considered sufficiently accurate. RESULTS: The mean age was 74.8 ± 8.6 years with a mean follow-up of 43.8 ± 22.1 months. Respectively, rates and risk factors of each outcome were 25.3% (n = 51) for abdominal aortic aneurysm (AAA) enlargement (age, number of patent sac branches, iliac calcifications and tortuosity, aneurysmal thrombus), 7% (n = 15) for type IA endoleak (neck calcification and AAA diameter), 3.7% (n = 8) for type IB endoleak (iliac tortuosity, AAA diameter, neck thrombus), 19.8% (n = 40) for type II endoleak (female, number of patent sac branches), and 25.9% (n = 55) for reintervention from any cause (neck calcification). The risk associated to each outcome can be calculated with a combination of these different preoperative variables. AUC for each outcome were 79.6% for AAA enlargement, 70.4% for reintervention, 81.3% for type IA endoleak, 92.3% for type IB endoleak, 70.6% for type II endoleak. CONCLUSIONS: This study shows that an exhaustive description of the preoperative anatomy before EVAR is a powerful and reliable tool to predict the risk of developing the most common complications after EVAR.

Endovascular aortic repair of a postdissecting thoracoabdominal aneurysm using intraoperative fusion imaging.

Computer-aided imaging can aid complex endovascular repair of aortic dissections in locating the narrow true lumen and identifying perfusion of visceral vessels by the true and/or false lumen. Although these anatomic data are available for analysis during planning, they are not readily available during the procedure with conventional imaging systems. We report the use of "fusion imaging" to facilitate the treatment of a postdissection thoracoabdominal aneurysm. The preoperative computer tomographic angiograms were processed, and the true and the false lumens were individually color labeled. These data were then superimposed on the fluoroscopic images in order to facilitate deployment of a fenestrated endograft.

Quantification of dose uncertainties in cumulated dose estimation compared to planned dose in prostate IMRT.

BACKGROUND AND PURPOSE: In prostate IMRT, the objectives were to quantify, for the bladder and the rectum: (i) the dose difference (DD) between the planned dose (PD) and the estimated cumulated dose (ECD) by deformable image registration (DIR); (ii) the dose accumulation uncertainties (DAU). MATERIALS AND METHODS: A series of 24 patients receiving 80Gy in the prostate was used to calculate the ECDpts and the DDpts. To evaluate the DAU, a numerical phantom (ph) simulating deformations of main pelvic organs was used to calculate the ECDph using the same DIR method. A reference cumulated dose (RCDph) was calculated, based on the simulated deformations. The DAUph was defined by the differences between RCDph and ECDph. RESULTS: For the mean dose to the bladder, the standard deviation of DDpts was 6.9Gy (18.1Gy maximum) with a DAUph of 2.7Gy. For the rectum wall, it was 2.0Gy (4.2Gy maximum) with a DAUph of 1.2Gy. Volume differences between PDpts and ECDpts, along the dose-volume histogram, ranged from -30% to +37% and -14% to +14% for the bladder and rectum, respectively. The corresponding uncertainties ranged from -23% to +7% and -4% to +7% for the bladder and rectum, respectively. CONCLUSIONS: Large differences between planned and delivered doses to the bladder have been quantified and are higher than the uncertainties of the DIR method. For the rectum, the differences are smaller and close to the DIR uncertainties.

Patient-specific numerical simulation of stent-graft deployment: Validation on three clinical cases.

Endovascular repair of abdominal aortic aneurysms faces some adverse outcomes, such as kinks or endoleaks related to incomplete stent apposition, which are difficult to predict and which restrain its use although it is less invasive than open surgery. Finite element simulations could help to predict and anticipate possible complications biomechanically induced, thus enhancing practitioners' stent-graft sizing and surgery planning, and giving indications on patient eligibility to endovascular repair. The purpose of this work is therefore to develop a new numerical methodology to predict stent-graft final deployed shapes after surgery. The simulation process was applied on three clinical cases, using preoperative scans to generate patient-specific vessel models. The marketed devices deployed during the surgery, consisting of a main body and one or more iliac limbs or extensions, were modeled and their deployment inside the corresponding patient aneurysm was simulated. The numerical results were compared to the actual deployed geometry of the stent-grafts after surgery that was extracted from postoperative scans. We observed relevant matching between simulated and actual deployed stent-graft geometries, especially for proximal and distal stents outside the aneurysm sac which are particularly important for practitioners. Stent locations along the vessel centerlines in the three simulations were always within a few millimeters to actual stents locations. This good agreement between numerical results and clinical cases makes finite element simulation very promising for preoperative planning of endovascular repair.

Finite element simulation of the insertion of guidewires during an EVAR procedure: example of a complex patient case, a first step toward patient-specific parameterized models.

Deformations of the vascular structure due to the insertion of tools during endovascular treatment of aneurysms of the abdominal aorta, unless properly anticipated during the preoperative planning phase, may be the source of intraoperative or postoperative complications. We propose here an explicit finite element simulation method which enables one to predict such deformations. This method is based on a mechanical model of the vascular structure which takes into account the nonlinear behavior of the arterial wall, the prestressing effect induced by the blood pressure and the mechanical support of the surrounding organs and structures. An analysis of the model sensitivity to the parameters used to represent this environment is done. This allows determining the parameters that have the largest influence on the quality of the prediction and also provides realistic values for each of them as no experimental data are available in the literature. Moreover, for the first time, the results are compared with 3D intraoperative data. This is done for a patient-specific case with a complex anatomy in order to assess the feasibility of the method. Finally, the predictive capability of the simulation is evaluated on a group of nine patients. The error between the final simulated and intraoperatively measured tool positions was 2.1 mm after the calibration phase on one patient. It results in a 4.6 ± 2.5 mm in average error for the blind evaluation on nine patients.

Surface-constrained nonrigid registration for dose monitoring in prostate cancer radiotherapy.

This paper addresses the issue of cumulative dose estimation from cone beam computed tomography (CBCT) images in prostate cancer radiotherapy. It focuses on the dose received by the surfaces of the main organs at risk, namely the bladder and rectum. We have proposed both a surface-constrained dose accumulation approach and its extensive evaluation. Our approach relied on the nonrigid registration (NRR) of daily acquired CBCT images on the planning CT image. This proposed NRR method was based on a Demons-like algorithm, implemented in combination with mutual information metric. It allowed for different levels of geometrical constraints to be considered, ensuring a better point to point correspondence, especially when large deformations occurred, or in high dose gradient areas. The three following implementations: 1) full iconic NRR; 2) iconic NRR constrained with landmarks (LCNRR); 3) NRR constrained with full delineation of organs (DBNRR). To obtain reference data, we designed a numerical phantom based on finite-element modeling and image simulation. The methods were assessed on both the numerical phantom and real patient data in order to quantify uncertainties in terms of dose accumulation. The LCNRR method appeared to constitute a good compromise for dose monitoring in clinical practice.

Finite-element-based matching of pre- and intraoperative data for image-guided endovascular aneurysm repair.

Endovascular repair of abdominal aortic aneurysms is a well-established technique throughout the medical and surgical communities. Although increasingly indicated, this technique does have some limitations. Because intervention is commonly performed under fluoroscopic control, 2-D visualization of the aneurysm requires the injection of a contrast agent. The projective nature of this imaging modality inevitably leads to topographic errors, and does not give information on arterial wall quality at the time of deployment. A specially adapted intraoperative navigation interface could increase deployment accuracy and reveal such information, which preoperative 3-D imaging might otherwise provide. One difficulty is the precise matching of preoperative data (images and models) and intraoperative observations affected by anatomical deformations due to tool-tissue interactions. Our proposed solution involves a finite-element-based preoperative simulation of tool-tissue interactions, its adaptive tuning regarding patient specific data, and the matching with intraoperative data. The biomechanical model was first tuned on a group of ten patients and assessed on a second group of eight patients.

Prediction of deformations during endovascular aortic aneurysm repair using finite element simulation.

During endovascular aortic aneurysm repair (EVAR), the introduction of medical devices deforms the arteries. The aim of the present study was to assess the feasibility of finite element simulation to predict arterial deformations during EVAR. The aortoiliac structure was extracted from the preoperative CT angiography of fourteen patients underwent EVAR. The simulation consists in modeling the deformation induced by the stiff wire used during EVAR. The results of the simulation were projected onto the intraoperative images, using a 3D/2D registration. The mean distance between the real and simulated guidewire was 2.3±1.1mm. Our results demonstrate that finite element simulation is feasible and appear to be reproducible in modeling device/tissue interactions and quantifying anatomic deformations during EVAR.