How to measure cortical folding from MR images: a step-by-step tutorial to compute local gyrification index.

Cortical folding (gyrification) is determined during the first months of life, so that adverse events occurring during this period leave traces that will be identifiable at any age. As recently reviewed by Mangin and colleagues(2), several methods exist to quantify different characteristics of gyrification. For instance, sulcal morphometry can be used to measure shape descriptors such as the depth, length or indices of inter-hemispheric asymmetry(3). These geometrical properties have the advantage of being easy to interpret. However, sulcal morphometry tightly relies on the accurate identification of a given set of sulci and hence provides a fragmented description of gyrification. A more fine-grained quantification of gyrification can be achieved with curvature-based measurements, where smoothed absolute mean curvature is typically computed at thousands of points over the cortical surface(4). The curvature is however not straightforward to comprehend, as it remains unclear if there is any direct relationship between the curvedness and a biologically meaningful correlate such as cortical volume or surface. To address the diverse issues raised by the measurement of cortical folding, we previously developed an algorithm to quantify local gyrification with an exquisite spatial resolution and of simple interpretation. Our method is inspired of the Gyrification Index(5), a method originally used in comparative neuroanatomy to evaluate the cortical folding differences across species. In our implementation, which we name local Gyrification Index (lGI(1)), we measure the amount of cortex buried within the sulcal folds as compared with the amount of visible cortex in circular regions of interest. Given that the cortex grows primarily through radial expansion(6), our method was specifically designed to identify early defects of cortical development. In this article, we detail the computation of local Gyrification Index, which is now freely distributed as a part of the FreeSurfer Software (http://surfer.nmr.mgh.harvard.edu/, Martinos Center for Biomedical Imaging, Massachusetts General Hospital). FreeSurfer provides a set of automated reconstruction tools of the brain's cortical surface from structural MRI data. The cortical surface extracted in the native space of the images with sub-millimeter accuracy is then further used for the creation of an outer surface, which will serve as a basis for the lGI calculation. A circular region of interest is then delineated on the outer surface, and its corresponding region of interest on the cortical surface is identified using a matching algorithm as described in our validation study(1). This process is repeatedly iterated with largely overlapping regions of interest, resulting in cortical maps of gyrification for subsequent statistical comparisons (Fig. 1). Of note, another measurement of local gyrification with a similar inspiration was proposed by Toro and colleagues(7), where the folding index at each point is computed as the ratio of the cortical area contained in a sphere divided by the area of a disc with the same radius. The two implementations differ in that the one by Toro et al. is based on Euclidian distances and thus considers discontinuous patches of cortical area, whereas ours uses a strict geodesic algorithm and include only the continuous patch of cortical area opening at the brain surface in a circular region of interest.

The Foundational Significance of Leggett's Non-local Hidden-Variable Theories

Laudisa (Found. Phys. 38:1110-1132, 2008) claims that experimental research on the class of non-local hidden-variable theories introduced by Leggett is misguided, because these theories are irrelevant for the foundations of quantum mechanics. I show that Laudisa's arguments fail to establish the pessimistic conclusion he draws from them. In particular, it is not the case that Leggett-inspired research is based on a mistaken understanding of Bell's theorem, nor that previous no-hidden-variable theorems already exclude Leggett's models. Finally, I argue that the framework of Bohmian mechanics brings out the importance of Leggett tests, rather than proving their irrelevance, as Laudisa supposes.

Novel anesthetics and other treatment strategies for refractory status epilepticus

Refractory status epilepticus (RSE)-that is, seizures resistant to at least two antiepileptic drugs (AEDs)-is generally managed with barbiturates, propofol, or midazolam, despite a low level of evidence (Rossetti, 2007). When this approach fails, the need for alternative pharmacologic and nonpharmacologic strategies emerges. These have been investigated even less systematically than the aforementioned compounds, and are often used, sometimes in succession, in cases of extreme refractoriness (Robakis & Hirsch, 2006). Several possibilities are reviewed here. In view of the marked heterogeneity of reported information, etiologies, ages, and comedications, it is extremely difficult to evaluate a given method, not to say to compare different strategies among them. Pharmacologic Approaches Isoflurane and desflurane may complete the armamentarium of anesthetics,' and should be employed in a ''close'' environment, in order to prevent intoxication of treating personnel. c-Aminobutyric acid (GABA)A receptor potentiation represents the putative mechanism of action. In an earlier report, isoflurane was used for up to 55 h in nine patients, controlling seizures in all; mortality was, however, 67% (Kofke et al., 1989). More recently, the use of these inhalational anesthetics was described in seven subjects with RSE, for up to 26 days, with an endtidal concentration of 1.2-5%. All patients required vasopressors, and paralytic ileus occurred in three; outcome was fatal in three patients (43%) (Mirsattari et al., 2004). Ketamine, known as an emergency anesthetic because of its favorable hemodynamic profile, is an N-methyl-daspartate (NMDA) antagonist; the interest for its use in RSE derives from animal works showing loss of GABAA efficacy and maintained NMDA sensitivity in prolonged status epilepticus (Mazarati & Wasterlain, 1999). However, to avoid possible neurotoxicity, it appears safer to combine ketamine with GABAergic compounds (Jevtovic-Todorovic et al., 2001; Ubogu et al., 2003), also because of a likely synergistic effect (Martin & Kapur, 2008). There are few reported cases in humans, describing progressive dosages up to 7.5 mg/kg/h for several days (Sheth & Gidal, 1998; Quigg et al., 2002; Pruss & Holtkamp, 2008), with moderate outcomes. Paraldehyde acts through a yet-unidentified mechanism, and appears to be relatively safe in terms of cardiovascular tolerability (Ramsay, 1989; Thulasimani & Ramaswamy, 2002), but because of the risk of crystal formation and its reactivity with plastic, it should be used only as fresh prepared solution in glass devices (Beyenburg et al., 2000). There are virtually no recent reports regarding its use in adults RSE, whereas rectal paraldehyde in children with status epilepticus resistant to benzodiazepines seems less efficacious than intravenous phenytoin (Chin et al., 2008). Etomidate is another anesthetic agent for which the exact mechanism of action is also unknown, which is also relatively favorable regarding cardiovascular side effects, and may be used for rapid sedation. Its use in RSE was reported in eight subjects (Yeoman et al., 1989). After a bolus of 0.3 mg/kg, a drip of up to 7.2 mg/kg/h for up to 12 days was administered, with hypotension occurring in five patients; two patients died. A reversible inhibition of cortisol synthesis represents an important concern, limiting its widespread use and implying a careful hormonal substitution during treatment (Beyenburg et al., 2000). Several nonsedating approaches have been reported. The use of lidocaine in RSE, a class Ib antiarrhythmic agent modulating sodium channels, was reviewed in 1997 (Walker & Slovis, 1997). Initial boluses up to 5 mg/kg and perfusions of up to 6 mg/kg/h have been mentioned; somewhat surprisingly, at times lidocaine seemed to be successful in controlling seizures in patients who were refractory to phenytoin. The aforementioned dosages should not be overshot, in order to keep lidocaine levels under 5 mg/L and avoid seizure induction (Hamano et al., 2006). A recent pediatric retrospective survey on 57 RSE episodes (37 patients) described a response in 36%, and no major adverse events; mortality was not given (Hamano et al., 2006 Verapamil, a calcium-channel blocker, also inhibits P-glycoprotein, a multidrug transporter that may diminish AED availability in the brain (Potschka et al., 2002). Few case reports on its use in humans are available; this medication nevertheless appears relatively safe (under cardiac monitoring) up to dosages of 360 mg/day (Iannetti et al., 2005). Magnesium, a widely used agent for seizures elicited by eclampsia, has also been anecdotally reported in RSE (Fisher et al., 1988; Robakis & Hirsch, 2006), but with scarce results even at serum levels of 14 mm. The rationale may be found in the physiologic blockage of NMDA channels by magnesium ions (Hope & Blumenfeld, 2005). Ketogenic diet has been prescribed for decades, mostly in children, to control refractory seizures. Its use in RSE as ''ultima ratio'' has been occasionally described: three of six children (Francois et al., 2003) and one adult (Bodenant et al., 2008) were responders. This approach displays its effect subacutely over several days to a few weeks. Because ''malignant RSE'' seems at times to be the consequence of immunologic processes (Holtkamp et al., 2005), a course of immunomodulatory treatment is often advocated in this setting, even in the absence of definite autoimmune etiologies (Robakis & Hirsch, 2006); steroids, adrenocorticotropic hormone (ACTH), plasma exchanges, or intravenous immunoglobulins may be used alone or in sequential combination. Nonpharmacologic Approaches These strategies are described somewhat less frequently than pharmacologic approaches. Acute implantation of vagus nerve stimulation (VNS) has been reported in RSE (Winston et al., 2001; Patwardhan et al., 2005; De Herdt et al., 2009). Stimulation was usually initiated in the operation room, and intensity progressively adapted over a few days up to 1.25 mA (with various regimens regarding the other parameters), allowing a subacute seizure control; one transitory episode of bradycardia/asystole has been described (De Herdt et al., 2009). Of course, pending identification of a definite seizure focus, resective surgery may also be considered in selected cases (Lhatoo & Alexopoulos, 2007). Low-frequency (0.5 Hz) transcranial magnetic stimulation (TMS) at 90% of the resting motor threshold has been reported to be successful for about 2 months in a patient with epilepsia partialis continua, but with a weaning effect afterward, implying the need for a repetitive use (Misawa et al., 2005). More recently, TMS was applied in a combination of a short ''priming'' high frequency (up to 100 Hz) and longer runs of low-frequency stimulations (1 Hz) at 90-100% of the motor threshold in seven other patients with simple-partial status, with mixed results (Rotenberg et al., 2009). Paradoxically at first glance, electroconvulsive treatment may be found in cases of extremely resistant RSE. A recent case report illustrates its use in an adult patient with convulsive status, with three sessions (three convulsions each) carried out over 3 days, resulting in a moderate recovery; the mechanism is believed to be related to modification of the synaptic release of neurotransmitters (Cline & Roos, 2007). Therapeutic hypothermia, which is increasingly used in postanoxic patients (Oddo et al., 2008), has been the object of a recent case series in RSE (Corry et al., 2008). Reduction of energy demand, excitatory neurotransmission, and neuroprotective effects may account for the putative mechanism of action. Four adult patients in RSE were cooled to 31_-34_C with an endovascular system for up to 90 h, and then passively rewarmed over 2-50 h. Seizures were controlled in two patients, one of whom died; also one of the other two patients in whom seizures continued subsequently deceased. Possible side effects are related to acid-base and electrolyte disturbances, and coagulation dysfunction including thrombosis, infectious risks, cardiac arrhythmia, and paralytic ileus (Corry et al., 2008; Cereda et al., 2009). Finally, anecdotic evidence suggests that cerebrospinal fluid (CSF)-air exchange may induce some transitory benefit in RSE (Kohrmann et al., 2006); although this approach was already in use in the middle of the twentieth century, the mechanism is unknown. Acknowledgment A wide spectrum of pharmacologic (sedating and nonsedating) and nonpharmacologic (surgical, or involving electrical stimulation) regimens might be applied to attempt RSE control. Their use should be considered only after refractoriness to AED or anesthetics displaying a higher level of evidence. Although it seems unlikely that these uncommon and scarcely studied strategies will influence the RSE outcome in a decisive way, some may be interesting in particular settings. However, because the main prognostic determinant in status epilepticus appears to be related to the underlying etiology rather than to the treatment approach (Rossetti et al., 2005, 2008), the safety issue should always represent a paramount concern for the prescribing physician. Conclusion The author confirms that he has read the Journal's position on issues involved in ethical publication and affirms that this paper is consistent with those guidelines.

CPT-11 and concomitant hyperfractionated accelerated radiotherapy induce efficient local control in rectal cancer patients: results from a phase II.

Patients with rectal cancer are at high risk of disease recurrence despite neoadjuvant radiochemotherapy with 5-Fluorouracil (5FU), a regimen that is now widely applied. In order to develop a regimen with increased antitumour activity, we previously established the recommended dose of neoadjuvant CPT-11 (three times weekly 90 mg m(-2)) concomitant to hyperfractionated accelerated radiotherapy (HART) followed by surgery within 1 week. Thirty-three patients (20 men) with a locally advanced adenocarcinoma of the rectum were enrolled in this prospective phase II trial (1 cT2, 29 cT3, 3 cT4 and 21 cN+). Median age was 60 years (range 43-75 years). All patients received all three injections of CPT-11 and all but two patients completed radiotherapy as planned. Surgery with total mesorectal excision (TME) was performed within 1 week (range 2-15 days). The preoperative chemoradiotherapy was overall well tolerated, 24% of the patients experienced grade 3 diarrhoea that was easily manageable. At a median follow-up of 2 years no local recurrence occurred, however, nine patients developed distant metastases. The 2-year disease-free survival was 66% (95% confidence interval 0.48-0.83). Neoadjuvant CPT-11 and HART allow for excellent local control; however, distant relapse remains a concern in this patient population.

Alterations in the local myocardial motion pattern in patients suffering from pressure overload due to aortic stenosis.

BACKGROUND: MR tissue tagging allows the noninvasive assessment of the locally and temporally resolved motion pattern of the left ventricle. Alterations in cardiac torsion and diastolic relaxation of the left ventricle were studied in patients with aortic stenosis and were compared with those of healthy control subjects and championship rowers with physiological volume-overload hypertrophy. METHODS AND RESULTS: Twelve aortic stenosis patients, 11 healthy control subjects with normal left ventricular function, and 11 world-championship rowers were investigated for systolic and diastolic heart wall motion on a basal and an apical level of the myocardium. Systolic torsion and untwisting during diastole were examined by use of a novel tagging technique (CSPAMM) that provides access to systolic and diastolic motion data. In the healthy heart, the left ventricle performs a systolic wringing motion, with a counterclockwise rotation at the apex and a clockwise rotation at the base. Apical untwisting precedes diastolic filling. In the athlete's heart, torsion and untwisting remain unchanged compared with those of the control subjects. In aortic stenosis patients, torsion is significantly increased and diastolic apical untwisting is prolonged compared with those of control subjects or athletes. CONCLUSIONS: Torsional behavior as observed in pressure- and volume-overloaded hearts is consistent with current theoretical findings. A delayed diastolic untwisting in the pressure-overloaded hearts of the patients may contribute to a tendency toward diastolic dysfunction.

Ulnar nerve transposition at the elbow : local anesthesia should be preferred to general anesthesia

Après la compression du nerf médian au niveau du tunnel carpien, la compression du nerf ulnaire au niveau du coude est le deuxième syndrome compressif le plus fréquent des nerfs périphériques. La chirurgie des nerfs périphériques consiste dans une décompression nerveuse et est caractérisée par un suivi post¬opératoire parfois très douloureux avec des douleurs qui peuvent chroniciser si insuffisamment traitées. Le traitement chirurgical de décompression nerveuse se fait traditionnellement sous anesthésie générale ou régionale. Une analgésie post-opératoire plus efficace et durable avec moindre risque de chronicisation avait justifié ce choix jusqu'à ce jour. Grâce au développement de la chirurgie ambulatoire ces dernières années, un grand nombre d'interventions chirurgicales au niveau de la main sont effectués sous anesthésie locale. Au vu d'une meilleure connaissance de cette technique d'anesthésie, son rôle dans la chirurgie des nerfs périphériques a été remis en question. Alors que plusieurs études ont démontré que l'anesthésie locale est aussi efficace que l'anesthésie générale et régionale au sujet de la chirurgie du tunnel carpien, son utilisation pour la chirurgie du nerf ulnaire reste peu connue. La raison de l'hésitation à l'utilisation de l'anesthésie locale pour le traitement du tunnel ulnaire est due au fait que dans plus de la moitié des cas, une simple décompression n'est pas suffisante et qu'il est souvent nécessaire de transposer le nerf ulnaire devant l'épicondyle ulnaire. La seule publication disponible au sujet de l'utilisation de l'anesthésie locale dans le traitement du tunnel ulnaire considère comme irréalisable d'utiliser cette méthode dans le cas d'une transposition. Malgré cette mise en garde, nous avons, depuis plusieurs années, des excellents résultats avec la transposition du nerf ulnaire sous anesthésie locale. Avec le but d'objectiver notre expérience dans ce domaine nous avons souhaité analyser nos résultats de façon rétrospective avec particulière attention aux douleurs post-opératoires et à la satisfaction des patients. Les dossiers de cinquante patients Consécutifs (26F, 24M) opérés par le même chirurgien dans notre service de 2002 à 2012 ont été analysés rétrospectivement. Les critères suivants ont été récoltés: l'âge du patient, la profession, la main dominante, les détails des techniques opératoires utilisées, le type d'anesthésie, l'intensité du suivi ainsi que les complications, le niveau de douleur dans l'immédiat post-opératorie ainsi que à une année de l'intervention. Les patients ont été divisés en 4 groupes: les opérés sous anesthésie générale avec transposition du nerf (n=17) ou sans transposition (n=10) et les opérés sous anesthésie locale avec transposition (n=12) ou sans transposition (n=11). Au premier jour la douleur était comparable dans tous les groupes. A une semaine, elle était deux fois plus importante lorsque la transposition avait été réalisée sous anesthésie générale par rapport à si une anesthésie locale avait été effectuée (p=0.03). La satisfaction s'est révélée plus élevée mais non significative chez les patients opérés sous anesthésie locale. Ces derniers étaient significativement plus enclins à répéter la chirurgie comparé a ceux opérés sous anesthésie générale (p=0.04). En conclusion, les résultats de cette étude suggèrent que l'anesthésie locale est au moins autant efficace que l'anesthésie générale en termes de complications et de douleurs post-opératoires indépendamment du fait qu'une transposition nerveuse soit effectuée ou pas. Un meilleur contrôle des douleurs à une semaine post-opératoire a contribué à une haute satisfaction des patients de notre étude.

Mass movement characterization using a reflexion and refraction seismic survey with the sloping local base level concept

This study proposes a new concept for upscaling local information on failure surfaces derived from geophysical data, in order to develop the spatial information and quickly estimate the magnitude and intensity of a landslide. A new vision of seismic interpretation on landslides is also demonstrated by taking into account basic geomorphic information with a numeric method based on the Sloping Local Base Level (SLBL). The SLBL is a generalization of the base level defined in geomorphology applied to landslides, and allows the calculation of the potential geometry of the landslide failure surface. This approach was applied to a large scale landslide formed mainly in gypsum and situated in a former glacial valley along the Rhone within the Western European Alps. Previous studies identified the existence of two sliding surfaces that may continue below the level of the valley. In this study. seismic refraction-reflexion surveys were carried out to verify the existence of these failure surfaces. The analysis of the seismic data provides a four-layer model where three velocity layers (<1000 ms(-1), 1500 ms(-1) and 3000 ms(-1)) are interpreted as the mobilized mass at different weathering levels and compaction. The highest velocity layer (>4000 ms(-1)) with a maximum depth of similar to 58 m is interpreted as the stable anhydrite bedrock. Two failure surfaces were interpreted from the seismic surveys: an upper failure and a much deeper one (respectively 25 and 50 m deep). The upper failure surface depth deduced from geophysics is slightly different from the results obtained using the SLBL, and the deeper failure surface depth calculated with the SLBL method is underestimated in comparison with the geophysical interpretations. Optimal results were therefore obtained by including the seismic data in the SLBL calculations according to the geomorphic limits of the landslide (maximal volume of mobilized mass = 7.5 x 10(6) m(3)).

Effect of low-dose chest irradiation before chemotherapy on the rate of local failure in small-cell lung cancer.

PURPOSE: To evaluate the rate of tumor recurrence within the irradiated volume after initial low-dose irradiation of limited-stage small-cell lung cancer (SCLC), to assess the tolerance of a sequential combination of low-dose chest irradiation followed by chemotherapy, and to confirm the responsiveness of limited-stage SCLC to low-dose irradiation. METHODS AND MATERIALS: In this pilot study, 26 patients with limited-stage SCLC were treated by first-line 20-Gy thoracic irradiation followed 3 weeks later by chemotherapy (cisplatin, doxorubicin, and etoposide for six cycles). RESULTS: We present our final results with a median follow-up of surviving patients of 7 years. The response rate to this low-dose irradiation was 83%, with an overall response rate to radiochemotherapy of 96% and a median survival of 21 months. No unexpected early or late toxicity was observed. The rate of initial isolated local failure was 8%, which compares favorably with other published series using higher doses of radiochemotherapy. CONCLUSION: An initial chest irradiation of 20 Gy before chemotherapy could be sufficient to reduce the risk of local failure during the time of survival of patients with limited-stage SCLC. Potential advantages of this treatment may be the prevention of resistance mechanisms to radiotherapy induced by preliminary chemotherapy and a reduced radiation-induced toxicity.

Integration of local-scale hydrological and regional-scale geophysical based on a nonlinear Bayesian sequential simulation approach

Significant progress has been made with regard to the quantitative integration of geophysical and hydrological data at the local scale. However, extending the corresponding approaches to the regional scale represents a major, and as-of-yet largely unresolved, challenge. To address this problem, we have developed a downscaling procedure based on a non-linear Bayesian sequential simulation approach. The basic objective of this algorithm is to estimate the value of the sparsely sampled hydraulic conductivity at non-sampled locations based on its relation to the electrical conductivity, which is available throughout the model space. The in situ relationship between the hydraulic and electrical conductivities is described through a non-parametric multivariate kernel density function. This method is then applied to the stochastic integration of low-resolution, re- gional-scale electrical resistivity tomography (ERT) data in combination with high-resolution, local-scale downhole measurements of the hydraulic and electrical conductivities. Finally, the overall viability of this downscaling approach is tested and verified by performing and comparing flow and transport simulation through the original and the downscaled hydraulic conductivity fields. Our results indicate that the proposed procedure does indeed allow for obtaining remarkably faithful estimates of the regional-scale hydraulic conductivity structure and correspondingly reliable predictions of the transport characteristics over relatively long distances.

A new method to estimate the infilling of alluvial sediment of glacial valleys using a sloping local base level

A new method is used to estimate the volumes of sediments of glacial valleys. This method is based on the concept of sloping local base level and requires only a digital terrain model and the limits of the alluvial valleys as input data. The bedrock surface of the glacial valley is estimated by a progressive excavation of the digital elevation model (DEM) of the filled valley area. This is performed using an iterative routine that replaces the altitude of a point of the DEM by the mean value of its neighbors minus a fixed value. The result is a curved surface, quadratic in 2D. The bedrock surface of the Rhone Valley in Switzerland was estimated by this method using the free digital terrain model Shuttle Radar Topography Mission (SRTM) (~92 m resolution). The results obtained are in good agreement with the previous estimations based on seismic profiles and gravimetric modeling, with the exceptions of some particular locations. The results from the present method and those from the seismic interpretation are slightly different from the results of the gravimetric data. This discrepancy may result from the presence of large buried landslides in the bottom of the Rhone Valley.

Inhibition of the Kit Ligand/c-Kit Axis Attenuates Metastasis in a Mouse Model Mimicking Local Breast Cancer Relapse after Radiotherapy.

PURPOSE: Local breast cancer relapse after breast-saving surgery and radiotherapy is associated with increased risk of distant metastasis formation. The mechanisms involved remain largely elusive. We used the well-characterized 4T1 syngeneic, orthotopic breast cancer model to identify novel mechanisms of postradiation metastasis. EXPERIMENTAL DESIGN: 4T1 cells were injected in 20 Gy preirradiated mammary tissue to mimic postradiation relapses, or in nonirradiated mammary tissue, as control, of immunocompetent BALB/c mice. Molecular, biochemical, cellular, histologic analyses, adoptive cell transfer, genetic, and pharmacologic interventions were carried out. RESULTS: Tumors growing in preirradiated mammary tissue had reduced angiogenesis and were more hypoxic, invasive, and metastatic to lung and lymph nodes compared with control tumors. Increased metastasis involved the mobilization of CD11b(+)c-Kit(+)Ly6G(high)Ly6C(low)(Gr1(+)) myeloid cells through the HIF1-dependent expression of Kit ligand (KitL) by hypoxic tumor cells. KitL-mobilized myeloid cells homed to primary tumors and premetastatic lungs, to give rise to CD11b(+)c-Kit(-) cells. Pharmacologic inhibition of HIF1, silencing of KitL expression in tumor cells, and inhibition of c-Kit with an anti-c-Kit-blocking antibody or with a tyrosine kinase inhibitor prevented the mobilization of CD11b(+)c-Kit(+) cells and attenuated metastasis. C-Kit inhibition was also effective in reducing mobilization of CD11b(+)c-Kit(+) cells and inhibiting lung metastasis after irradiation of established tumors. CONCLUSIONS: Our work defines KitL/c-Kit as a previously unidentified axis critically involved in promoting metastasis of 4T1 tumors growing in preirradiated mammary tissue. Pharmacologic inhibition of this axis represents a potential therapeutic strategy to prevent metastasis in breast cancer patients with local relapses after radiotherapy. Clin Cancer Res; 18(16); 4365-74. ©2012 AACR.