Deep Learning Classification of 3.5 GHz Band Spectrograms with Applications to Spectrum Sensing.

In the United States, the Federal Communications Commission has adopted rules permitting commercial wireless networks to share spectrum with federal incumbents in the 3.5 GHz Citizens Broadband Radio Service band. These rules require commercial systems to vacate the band when sensors detect radars operated by the U.S. military; a key example being the SPN-43 air traffic control radar. Such sensors require highly-accurate detection algorithms to meet their operating requirements. In this paper, using a library of over 14,000 3.5 GHz band spectrograms collected by a recent measurement campaign, we investigate the performance of thirteen methods for SPN-43 radar detection. Namely, we compare classical methods from signal detection theory and machine learning to several deep learning architectures. We demonstrate that machine learning algorithms appreciably outperform classical signal detection methods. Specifically, we find that a three-layer convolutional neural network offers a superior tradeoff between accuracy and computational complexity. Last, we apply this three-layer network to generate descriptive statistics for the full 3.5 GHz spectrogram library. Our findings highlight potential weaknesses of classical methods and strengths of modern machine learning algorithms for radar detection in the 3.5 GHz band.

The ultrafast reactions in the photochromic cycle of water-soluble fulgimide photoswitches.

Photochromic switches are essential for the control and manipulation of nanoscale reactions and processes. The expansion of their application to aqueous environments depends strongly on the development of optimized water-soluble photoswitches. Here we present a femtosecond time-resolved investigation of the photochromic reactions (transition between the open and the closed form) of a water-soluble indolylfulgimide. We observe a pronounced effect of the protic nature of water as a solvent on the ultrafast ring-opening reaction. Typically, the excited state of the closed form has a larger dipole moment than the ground state, which leads to stabilization of the excited state in polar solvents and hence a lifetime (3 ps) longer than in non-polar solvents (2 ps). However, in water, despite the increased solvent polarity and the increased excited state dipole moment, the opposite trend for the excited state lifetime is observed (1.8 ps). This effect is caused by the opening of a new excited state deactivation pathway involving proton transfer reactions.

Early indicators of treatment success after percutaneous radiofrequency of pulmonary tumors.

We retrospectively reviewed the imaging of patients after radiofrequency ablation (RFA) of lung metastases performed at our institution to assess the usefulness of ground glass opacification (GGO) margin for the prediction of complete tumor ablation. From January 2004 to March 2007, patients were identified where there was a postprocedure thin collimation scan to allow multiplanar reformatting, either immediately or at 24 h and at least 6 months of imaging follow-up. Thirty-six tumors in 22 patients were identified. The scans were assessed for the presence and width of GGO margin, and minimal and maximal dimensions were measured. A second reviewer, blinded to the outcome of the postprocedure assessment, reviewed the follow-up imaging for recurrence. The recurrence group had larger tumors (p = 0.045) and smaller mean minimal GGO margin width (p = 0.0001). Multivariate binary regression analysis confirmed that the minimal GGO margin was significantly (p < 0.005) associated with tumor recurrence. Receiver operator characteristic curve analysis suggests a cutoff of 4.5 mm for complete tumor ablation. There was substantial agreement (kappa = 0.759) between the site of absent GGO margin and the site of tumor recurrence. The point on the tumor surface where there is no GGO margin is likely to be the site of future recurrence. In our experience, a circumferential GGO margin of >5 mm is the minimal margion required to ensure complete tumor ablation.

Five-year survival in 309 patients with colorectal liver metastases treated with radiofrequency ablation.

There is little published long-term survival data for patients with colorectal liver metastases treated with radiofrequency ablation (RFA). We present a multivariate analysis of 5-year survival in 309 patients (198 male, aged 64 (24-92)) treated at 617 sessions. Our standard protocol used internally cooled electrodes introduced percutaneously under combined US and CT guidance/monitoring. The number and size of liver metastases, the presence and location of extrahepatic disease, primary resection, clinical, chemotherapy and follow-up data were recorded. Data analysis was performed using SPSS v.10. On multivariate analysis, significant survival factors were the presence of extrahepatic disease (p < 0.001) and liver tumour volume (p = 0.001). For 123 patients with five or less metastases of 5 cm or less maximum diameter and no extrahepatic disease median survival was 46 and 36 months from liver metastasis diagnosis and ablation, respectively; corresponding 3- and 5-year survival rates were 63%, 34% and 49%, 24%. Sixty-nine patients had three or less tumours of below 3.5 cm in diameter and their 5-year survival from ablation was 33%. There were 23/617(3.7%) local complications requiring intervention. Five-year survival of 24-33% post ablation in selected patients is superior to any published chemotherapy data and approaches the results of liver resection.

Quantitative secretin MRCP (MRCPQ): results in 215 patients with known or suspected pancreatic pathology.

Secretin magnetic resonance cholangiopancreatography quantification (MRCPQ) of pancreatic exocrine function correlates well with steatorrhoea and conventional, non-invasive function tests. We report MRCPQ results in a variety of pancreatic conditions. A total of 215 patients [107 male, mean age 46 years (14-78)] were studied. A multi-slice MRCP sequence was performed before and at 2-min intervals after 0.1 ml/kg IV secretin. Change in small intestinal water volume was plotted against time and the flow rate derived from the gradient. Patients were classified using clinical history, MRCP, MR imaging, computed tomography (CT) [150/215 (70%)] and endoscopic retrograde cholangiopancreatography (ERCP) [56/215 (26%)] findings but not MRCPQ results. Mean, standard deviation and 95% confidence intervals were calculated. The one way ANOVA and Student's t-test were used for statistical analysis. Seventy-six patients had chronic pancreatitis, 26 were post-surgical, 34 post-acute pancreatitis, six atrophic pancreatopathy, eight with obstruction, 15 divisum, ten sphincter of Oddi dysfunction, 26 normal and 14 miscellaneous. Significant differences were observed between normals (mean+/-SD; 7.4 +/- 2.9 ml/min) and severe chronic pancreatitis (5.3 +/- 2.4) (P = 0.018), pancreatic atrophy (3.8 +/- 3.1) (P = 0.013) or duct obstruction (5.3 +/- 2.4) (P = 0.047)) and between moderate (7.0 +/- 3.0) (P = 0.03) and severe chronic pancreatitis. MRCPQ can be used to quantify function across the spectrum of pancreatic disease and showed significant differences between several different pathologies.

Analysis of the factors associated with radiofrequency ablation-induced pneumothorax.

AIM: To define the characteristics most likely to result in radiofrequency ablation (RFA)-induced pneumothorax. METHODS AND MATERIALS: CT-guided RFA was performed in 79 tumours in 55 lungs in 37 patients, 16 were women, mean age 62 years (range 34-83). Three had primary lung cancer, 34 had metastases. The number, size, and location of tumours, electrode type, treatment parameters, length of electrode trajectory through aerated lung, background emphysema, prior interventions, and use of positive-pressure ventilation were analysed. The size, timing of any pneumothoraces, and intervention were recorded. RESULTS: Pneumothorax occurred in 21 of the 25 lungs treated (38%), 18 immediate and three delayed. Seventeen of the 21 (81%) occupied less than 30% of the hemithorax, whereas in four cases >31% was involved. Eight of the 55 (15%) pneumothoraces required aspiration. The length of the electrode trajectory through aerated lung in those who developed a pneumothorax was 5.4+/-4.7cm versus 1.9+/-2.7 in those who did not (p=0.001). The mean number of tumours ablated was higher in the pneumothorax group, 1.7+/-1 versus 1.3+/-0.6 (p=0.03), as was the number of electrode positions, 6+/-3.9 versus 3.6+/-2.2 (p=0.01). On multivariate analysis only the needle trajectory through aerated lung was significant (p=0.04). CONCLUSIONS: The number of tumours, electrode positions, and the anticipated electrode trajectory through aerated lung impacts on the likelihood of a pneumothorax. These considerations should be factored into patient selection, the choice of approach, and trajectory used in RFA.

Radiofrequency ablation (RFA) as a cytoreductive strategy for hepatic metastasis from breast cancer.

INTRODUCTION: Patients with liver metastasis from breast cancer have a poor prognosis, although this may be improved by hepatectomy in a selected group with disease confined to the liver. We evaluate the effectiveness of radiofrequency ablation (RFA) as a cytoreductive strategy in the management of liver metastasis from primary breast cancer. PATIENTS AND METHODS: Nineteen patients with hepatic metastasis from primary breast cancer underwent RFA of their liver lesions between April 1998 and August 2004. RESULTS: The median age of the patients was 52 years (range, 32-69 years), 8 had disease confined to the liver, with 11 having stable extrahepatic disease in addition. Seven patients with disease confined to the liver at presentation are alive, as are 6 with extrahepatic disease, median follow-up after RFA was 15 months (range, 0-77 months). Survival at 30 months was 41.6%. In addition, 7 patients followed up for a median of 14 months (range, 2-29 months) remain alive and disease-free. RFA failed to control hepatic disease in 3 patients. RFA was not associated with any mortality or major morbidity. CONCLUSIONS: Control of hepatic metastasis from breast cancer is possible using RFA and may lead to a survival benefit, particularly in those patients with disease confined to the liver.

Diagnosing pancreatic cancer: the role of percutaneous biopsy and CT.

AIMS: To determine the sensitivity and complications of percutaneous biopsy of pancreatic masses, and whether typical computed tomography (CT) features of adenocarcinoma can reliably predict this diagnosis. MATERIALS AND METHODS: A 5 year retrospective analysis of percutaneous core biopsies of pancreatic masses and their CT features was undertaken. Data were retrieved from surgical/pathology databases; medical records and CT reports and images. RESULTS: Three hundred and three patients underwent 372 biopsies; 56 of 87 patients had repeat biopsies. Malignancy was diagnosed in 276 patients, with ductal adenocarcinoma in 259 (85%). Final sensitivity of percutaneous biopsy for diagnosing pancreatic neoplasms was 90%; for repeat biopsy it was 87%. Complications occurred in 17 (4.6%) patients, in three of whom the complications were major (1%): one abscess, one duodenal perforation, one large retroperitoneal bleed. CT features typical of ductal adenocarcinoma were: hypovascular pancreatic mass with bile and/or pancreatic duct dilatation. Atypical CT features were: isodense or hypervascular mass, calcification, non-dilated ducts, cystic change, and extensive lymphadenopathy. Defining typical CT features of adenocarcinoma as true-positives, CT had a sensitivity of 68%, specificity of 95%, positive predictive value (PPV) of 98%, and negative predictive value of 41% for diagnosing pancreatic adenocarcinoma. CONCLUSION: Final sensitivity of percutaneous biopsy for establishing the diagnosis was 90%. CT features typical of pancreatic adenocarcinoma had high specificity and PPV. On some occasions, especially in frail patients with co-morbidity, it might be reasonable to assume a diagnosis of pancreatic cancer if CT features are typical, and biopsy only if CT shows atypical features.

Photodynamic therapy using meso tetra hydroxy phenyl chlorin (mTHPC) in early prostate cancer.

BACKGROUND AND OBJECTIVES: Prostate cancer is increasing in incidence, but current treatments including surgery and radiotherapy have significant side effects. This pilot study was designed to assess the potential of photodynamic therapy (PDT) using meso tetra hydroxy phenyl chlorin (mTHPC) for organ confined prostate cancer. STUDY DESIGN/PATIENTS AND METHODS: Six men with organ confined prostate cancer were photosensitised with mTHPC (0.15 mg/kg). Between 2 and 5 days later, red light (652 nm) was delivered to areas of biopsy proven cancer via fibres inserted through transperineal needles (50-100 J per site). RESULTS: After 8 of 10 PDT sessions, the prostate specific antigen (PSA) fell by up to 67%. Early MRI scans showed oedema and patchy necrosis, which resolved over 2 months. Biopsies of treated areas revealed necrosis and fibrosis at 1-2 months. CONCLUSIONS: PDT for primary prostate cancer appears safe and can reduce PSA levels. As this was a phase I study, no attempt was made to treat the whole prostate; this or targeted tumour ablation could be attempted in a phase II study with an increased number of fibres. This technique merits further investigation in early prostate cancer.

Diagnosis of duct disruption and assessment of pancreatic leak with dynamic secretin-stimulated MR cholangiopancreatography.

OBJECTIVE: The management of pancreatic duct disruption is complex and depends on several factors including the cause, morphology, and degree of disruption. ERCP can show duct disruption in as many as 75% of patients but is invasive and cannot detect disruption beyond an obstruction. We studied the role of secretin MR cholangiopancreatography in patients with suspected pancreatic duct disruption. CONCLUSION: Secretin MR cholangiopancreatography is a safe, noninvasive test that can provide additional useful information about duct integrity and facilitate management.

CT mapping of the distribution of saline during radiofrequency ablation with perfusion electrodes.

PURPOSE: During radiofrequency (RF) ablation, adjunctive saline increases the size of the ablation zone and therefore electrodes that simultaneously deliver current and saline have been developed, but the addition of saline also results in an irregular ablation zone. Our aim was to study the distribution of saline during RF ablation. METHODS: Four patients were treated: 3 with liver metastases and 1 with hepatocellular carcinoma (HCC). Two different perfusion electrodes were used: a high-perfusion-rate, straight electrode (Berchtold, Germany) and a low-perfusion-rate, expandable electrode (RITA Medical Systems, USA). The saline perfusate was doped with non-ionic contrast medium to render it visible on CT and the electrical conductivity was measured. CT scans were obtained of each electrode position prior to ablation and repeated after ablation. Contrast-enhanced CT was performed 18-24 hr later to demonstrate the ablation zone. All treatments were carried out according to the manufacturer's recommended protocol. RESULTS: The addition of a small quantity of non-ionic contrast did not alter the electrical conductivity of the saline. Contrast-doped saline extravasated beyond the tumor in all 3 patients with metastases but was limited in the patient with HCC. In some areas where saline had extravasated there was reduced enhancement on contrast-enhanced CT consistent with tissue ablation. One patient treated with the high-perfusion-rate system sustained a jejunal perforation requiring surgery. CONCLUSION: Saline can extravasate beyond the tumor and with the high-perfusion-rate system this resulted in an undesirable extension of the ablation zone and a complication.

Radiofrequency ablation of neuroendocrine liver metastases: the Middlesex experience.

BACKGROUND: Current treatment options for neuroendocrine liver metastases are not widely applicable or not that effective. Image-guided thermal ablation offers the possibility of a minimally invasive, albeit palliative, treatment that decreases tumor volume, preserves most of the normal liver, and can be repeated several times. We report our experience with image-guided thermal ablation in 25 patients with unresectable liver metastases. METHODS: Since 1990 we have treated 189 tumors at 66 treatment sessions in 25 patients (12 female, 13 male; median age, 56 years; age range, 26--78 years). Thirty treatments were performed with a solid-state laser, and 36 treatments were performed with radiofrequency ablation. All but one treatment was performed percutaneously under image guidance. Sixteen patients had metastases from carcinoid primaries, three from gastrinoma, two from insulinoma, and four from miscellaneous causes. Fourteen of 25 had symptoms from hormone secretion. RESULTS: Imaging follow-up was available in 19 patients at a median of 21 months (range, 4--75 months). There was a complete response in six patients, a partial response in seven, and stable disease in one; hence, tumor load was controlled in 14 of 19 patients (74%). Relief of hormone-related symptoms was achieved in nine of 14 patients (69%). The median survival period from the diagnosis of liver metastases was 53 months. One patient with end-stage cardiac disease died after a carcinoid crisis. There were eight (12%) complications: five local and three distant, four major and four minor. CONCLUSIONS: As a minimally invasive, readily repeatable procedure that can be used to ablate small tumors, preferably before patients become severely symptomatic, radiofrequency ablation can provide effective control of liver tumor volume in most patients over many years.

Radiofrequency ablation of colorectal liver metastases.

BACKGROUND: Untreated patients with colorectal liver metastases rarely survive 3 years, and the 3-year survival rate for patients treated with chemotherapy is 3%. The best survival rates are for the small subgroup that has operable disease, i.e., 39% at 5 years. Radiofrequency ablation (RFA) offers a new opportunity to destroy liver metastases in patients who are not surgical candidates because of disease distribution or comorbidity. METHODS: Acceptance criteria were a maximum of four or five liver lesions with a maximum diameter of 4 or 5 cm and no evidence of active extrahepatic disease. Nearly all treatments were performed percutaneously using ultrasound, computed tomography, or magnetic resonance imaging (or some combination) for guidance and monitoring. RFA is a minimally invasive procedure that can be readily repeated. General anesthesia facilitates the procedure but is not essential. Multiple overlapping ablations are required to ensure optimal treatment in all but the smallest tumors. RESULTS: In our cohort of 167 patients with colorectal liver metastases, 73 fulfilled the optimal acceptance criteria (5 or fewer tumors that were

Radio-frequency ablation of colorectal liver metastases in 167 patients.

The objective of this paper is to report our results from a prospective study of 167 patients with colorectal liver metastases treated with radio-frequency ablation (RFA). Three hundred fifty-four treatments were performed in 167 patients, 99 males, mean age 57 years (34-87). The mean number of metastases was 4.1 (1-27). The mean maximum diameter was 3.9 cm (1-12). Fifty-one (31%) had stable/treated extra-hepatic disease. Treatments were performed under general anaesthesia using US and CT guidance and single or cluster water-cooled electrodes (Valleylab, Boulder, CO). All patients had been rejected for or had refused surgical resection. Eighty percent received chemotherapy. Survival data were stratified by tumour burden at the time of first RFA. The mean number of RFA treatments was 2.1 (1-7). During a mean follow-up of 17 months (0-89), 72 developed new liver metastases and 71 developed progressive extra-hepatic disease. There were 14/354 (4%) major local complications and 22/354 (6%) minor local complications. For patients with < or =5 metastases, maximum diameter < or =5 cm and no extra-hepatic disease, the 5-year survival from the time of diagnosis was 30% and from the time of first thermal ablation was 26%. Given that the 5-year survival for operable patients is a median of 32%, our 5-year survival of 30% is promising.

Comparison of resection and radiofrequency ablation for treatment of solitary colorectal liver metastases.

BACKGROUND: Liver resection is the treatment of choice for patients with solitary colorectal liver metastases. In recent years, however, radiofrequency ablation has been used increasingly in the treatment of colorectal liver metastases. In the absence of randomized clinical trials, this study aimed to compare outcome in patients with solitary colorectal liver metastases treated by surgery or by radiofrequency ablation. METHODS: Solitary colorectal liver metastases were treated by radiofrequency destruction in 25 patients. The indications were extrahepatic disease in seven, vessel contiguity in nine and co-morbidity in nine patients. Outcome was compared with that of 20 patients who were treated by liver resection for solitary metastases and had no evidence of extrahepatic disease. Most patients in both groups also received systemic chemotherapy. RESULTS: Median survival after liver resection was 41 (range 0-97) months with a 3-year survival rate of 55.4 per cent. There was one postoperative death and morbidity was minimal. Median survival after radiofrequency ablation was 37 (range 9-67) months with a 3-year survival rate of 52.6 per cent. CONCLUSION: Survival after resection and radiofrequency ablation of solitary colorectal liver metastases was comparable. The latter is less invasive and requires either an overnight stay or day-case facilities only.

Radiofrequency ablation extends the scope of surgery in colorectal liver metastases.

AIMS: To assess outcome in patients treated by a multidisciplinary team, with a combination of liver resection and RF ablation. METHODS: Sixteen unselected patients (f=9; m=7) with colorectal liver metastases who were not suitable for surgery alone, were treated as follows: six had RF ablation at open laparotomy, three patients had synchronous ablation and resection while seven patients had RF ablation after liver resection. Standard liver resection techniques were used. RF was performed using internally cooled, single or cluster electrodes with a high power (200 W) generator. All patients were followed with regular contrast enhanced CT and survival noted. RESULTS: A total of 27 tumours with diameters 1.2-10 cm were treated. Two minor complications were recorded. 2/6 (33%) who had intraoperative RF had incomplete ablation due to large tumour size (6 and 10 cm respectively). Further RF ablation sessions were carried out successfully. 11/16 (69%) are alive at 2 years of whom 7 (44%) have no evidence of residual or recurrent liver disease. CONCLUSION: In our study, RF ablation extends the therapeutic envelope, is an effective local treatment of liver metastases and improves life expectancy.

Non-invasive quantification of pancreatic exocrine function using secretin-stimulated MRCP.

Our objective was to quantify water volume using magnetic resonance cholangiopancreatography (MRCP) sequences and apply this to secretin-stimulated studies with the aim of quantifying pancreatic exocrine function. A commercially available single-shot MRCP sequence was used in conjunction with a body phased-array coil and a 1.5-T MR system. Signal intensity was measured in samples of water, pancreatic, duodenal juice, and secretin-stimulated pancreatic juice. A water phantom was made and MR calculated volumes compared with known water volumes within the phantom. Changes in small intestinal volume in response to secretin were measured in a group of 11 patients with no evidence of pancreatic disease. Changes in water volume were plotted over time. The pancreatic duct diameter before and after secretin was noted and filling defects were sought. All patients also underwent an axial breath-hold T1-weighted gradient-echo sequence and the pancreatic parenchyma was evaluated for size and signal intensity. There was no difference in the signal intensity of the different juice samples. There was excellent correlation between known and calculated MRCP volumes (chi(2)=0.99). All patients demonstrated normal duct morphology on MRCP and normal pancreatic parenchyma on T1-weighted imaging. The mean flow rate in the patient population was 8.1+/-2.5 ml/s over a median of 7 min (range 5-9 min). The MRCP sequence can be used to measure water volume. Sequential MRCP measurements following secretin permitted calculation of volume change and flow rate. This should prove useful as an indicator of pancreatic exocrine function.

Comparison of B-mode and conventional colour/power Doppler ultrasound, contrast-enhanced Doppler ultrasound and spiral CT in the diagnosis of focal lesions of the liver: Results of a multicentre study.

AIM: In patients with lesions of the liver we compared diagnoses based on different methods of ultrasound as well as spiral CT with the final diagnosis reached at the time of the patient's discharge from hospital. METHOD: From records of a prospective multicentre study including 90 liver lesions investigated with B-mode baseline and conventional colour/power Doppler ultrasound, contrast-enhanced colour/power Doppler ultrasound and spiral-CT, we evaluated only those where diagnoses for all modalities were available, and where the diagnosis at discharge comprising all clinical, laboratory and imaging data as well as histologic proof was at least "highly probable". RESULTS: 60 lesions met the inclusion criteria. 20 lesions were ultimately diagnosed as benign, and 40 as malignant. With respect to the diagnosis of malignancy, sensitivity was 92.5 % (37/40) with B-mode and unenhanced conventional colour/power Doppler US ultrasound, 97.5 % (39/40) with contrast-enhanced colour/ power Doppler ultrasound, and 100 % with CT; the corresponding specificities were 65 % (13/20), 85 % (17/20) and 80 % (16/20). 4 of 7 false positive, and 2 of 3 false negative results in the unenhanced technique were diagnosed correctly with contrast-enhanced Doppler ultrasound. CONCLUSION: Compared to conventional ultrasound, contrast-enhanced Doppler ultrasound improved the diagnostic accuracy in 10 % of the cases. Its accuracy in our study was equal to that of CT.

Dynamic three-dimensional color Doppler ultrasound of human fetal intracardiac flow.

OBJECTIVES: To develop dynamic three-dimensional ultrasound techniques for prenatal imaging of the intracardiovascular flow as well as the cardiovascular structure to address difficulties in assessing the spatially complex hemodynamics and morphology of the fetal heart. METHODS: Gray-scale and color (velocity) Doppler echocardiography were performed on 12 fetuses to provide serial anatomical and rheological tomograms which were spatially registered in three dimensions. Using a second ultrasound machine simultaneously, spectral Doppler ultrasound was performed to record umbilical arterial waveforms, thus providing the temporal (fourth) dimension in terms of the cardiac cycle and facilitating removal of motion artifacts. RESULTS: Acquisitions were successful in eight of 15 attempts. Imaging of the flow of blood in four dimensions was achieved in six of the eight datasets. In one case with complex cardiac malformations, three-dimensional reconstructions at systole and diastole offered dynamic diagnostic views not appreciated on the cross-sectional images. CONCLUSIONS: Our novel technique has made possible the prenatal visualization of the spatial distribution and true direction of intracardiac flow of blood in four dimensions in the absence of motion artifacts. The technique suggests that diagnosis of cardiac malformations can be made on the basis of morphological and hemodynamic changes throughout the entire cardiac cycle, offering unique and significant information complementary to conventional techniques. Further work to integrate the several non-purpose-built machines into a single system will improve the rate of acquisition of data, and may provide a new means of imaging and modeling structure and hemodynamics, not only for the fetal heart but for many other moving body parts.