Endovascular stenting in the management of malignant superior vena cava obstruction: comparing safety, effectiveness, and outcomes between primary stenting and salvage stenting.

OBJECTIVE: To compare the safety, effectiveness, and outcomes of primary stenting and salvage stenting for malignant superior vena cava obstruction. DESIGN: Case series with internal comparison. SETTING: Regional hospital, Hong Kong. PATIENTS: A total of 56 patients with malignant superior vena cava obstruction underwent 59 stentings from 1 May 1999 to 31 January 2014. Patients' characteristics, procedural details, and outcomes were retrospectively reviewed. Of the 56 patients, 33 had primary stenting before conventional therapy and 23 had salvage stenting after failure of conventional therapy. Statistical analyses were made by Fisher's exact test and Mann-Whitney U test. RESULTS: Primary lung carcinoma was the most common cause of malignant superior vena cava obstruction (primary stenting, 22 patients; salvage stenting, 16 patients; P=0.768), followed by metastatic lymphadenopathy. Most patients had superior vena cava obstruction only (primary stenting, 16 patients; salvage stenting, 15 patients; P=0.633), followed by additional right brachiocephalic vein involvement. Wallstents (Boston Scientific, Natick [MA], US) were used in all patients. Technical success was achieved in all but two patients, one in each group (P=1.000). Only one stent placement was required in most patients (primary stenting, 28 patients; salvage stenting, 20 patients; P=0.726). Procedure time was comparable in both groups (mean time: primary stenting, 89 minutes; salvage stenting, 84 minutes; P=0.526). Symptomatic relief was achieved in most patients (primary stenting, 32 patients; salvage stenting, 23 patients; P=0.639). In-stent restenosis and bleeding were the commonest complications (primary stenting, 6 and 1 patients, respectively; salvage stenting, 2 and 2 patients, respectively). Nine patients required further treatment for symptom recurrence (primary stenting, 6 patients; salvage stenting, 3 patients; P=0.725). CONCLUSION: Endovascular stenting is safe and effective for relieving malignant superior vena cava obstruction. No statistically significant differences in number of stents, success rates, procedure times, symptom relief rates, complication rates, and re-procedure rates were found between primary stenting and salvage stenting.

Emergency transcatheter embolization of ruptured hepatocellular carcinomas with tortuous conventional or aberrant hepatic vascular anatomy, or parasitic supply.

Emergency transcatheter embolization is a well-recognized measure to manage patients with life-threatening haemoperitoneum due to spontaneous ruptured hepatocellular carcinoma. Traditional embolization technique is to embolize the proper hepatic artery or the segmental hepatic artery by femoral approach using gelfoam pledgets. From 1997 to 2004, in 19 out of 96 embolizations, the embolization technique had to be modified because of tortuous conventional or aberrant hepatic vascular anatomy or parasitic supply to achieve successful embolization.

Transcatheter embolisation of visceral pseudoaneurysms--technical difficulties and modification of embolisation technique.

We report 13 successful emergency embolisations for 12 pseudoaneurysms performed in 10 patients. For five pseudoaneurysms the embolisation technique was modified according to number of supplying vessels, flow rate of pseudoaneurysms, vascular anatomy and whether there was clinical evidence of re-bleeding or not. Apart from traditional embolisation technique, modified embolisation techniques are also useful for endovascular therapy of pseudoaneurysms.

Accuracy of nasopharyngeal carcinoma staging by magnetic resonance imaging.

To audit the accuracy of magnetic resonance (MR) staging of nasopharyngeal carcinoma (NPC) in daily reporting, the MR images of 101 adult patients with newly diagnosed NPC reported between December 1996 and February 2002 were reviewed retrospectively. Based on the American Joint Committee on Cancer/International Union Against Cancer (AJCC/UICC) TNM staging criteria (1997), the tumour staging obtained from the MR reports and the MR films was compared by two experienced head and neck radiologists. The number of patients being upstaged, downstaged or unchanged was noted. In all stages, the NPC staging obtained from the MR reports revealed 18 (17.8%) understaged, eight (7.9%) overstaged and 75 (74.2%) the same stage when compared to the staging obtained from the MR films based on the AJCC/UICC criteria. The percentage of patients being understaged or overstaged, in decreasing order of frequency, were stages II, III, IV and I. Magnetic resonance of NPC should be reported by radiologists who are not only familiar with the pathology of this condition and its pattern of spread but who should also base their reports on the AICC/UICC staging criteria. The necessity to improve our MR staging accuracy is largely because it determines the type of therapy to be given and this has to be appropriate and adequate for a successful treatment.

Embolisation of a renal artery pseudoaneurysm in a patient with renal malrotation and chronic aortic dissection.

INTRODUCTION: Renal artery pseudoaneurysms may arise as a complication of percutaneous nephrolithotomy (PCNL). Prompt recognition and treatment is essential to arrest haemorrhage which may be life threatening. CLINICAL PICTURE: A patient with chronic aortic dissection and malrotated right kidney underwent PCNL for right renal calculus. He developed delayed gross haematuria. TREATMENT: Angiography showed a pseudoaneurysm arising from one of two right renal arteries, which in turn arose from the false lumen of the aortic dissection. The supplying artery was successfully embolised. CONCLUSION: Renal artery pseudoaneurysms can be successfully treated with prompt angiography and embolisation, even in the presence of renal malrotation and aortic dissection.

An adaptive Bayesian pruning for neural networks in a non-stationary environment.

Pruning a neural network to a reasonable smaller size, and if possible to give a better generalization, has long been investigated. Conventionally the common technique of pruning is based on considering error sensitivity measure, and the nature of the problem being solved is usually stationary. In this article, we present an adaptive pruning algorithm for use in a nonstationary environment. The idea relies on the use of the extended Kalman filter (EKF) training method. Since EKF is a recursive Bayesian algorithm, we define a weight-importance measure in term of the sensitivity of a posteriori probability. Making use of this new measure and the adaptive nature of EKF, we devise an adaptive pruning algorithm called adaptive Bayesian pruning. Simulation results indicate that in a noisy nonstationary environment, the proposed pruning algorithm is able to remove network redundancy adaptively and yet preserve the same generalization ability.

Analysis for a class of winner-take-all model.

Recently we have proposed a simple circuit of winner-take-all (WTA) neural network. Assuming no external input, we have derived an analytic equation for its network response time. In this paper, we further analyze the network response time for a class of winner-take-all circuits involving self-decay and show that the network response time of such a class of WTA is the same as that of the simple WTA model.

On the Kalman filtering method in neural network training and pruning.

In the use of extended Kalman filter approach in training and pruning a feedforward neural network, one usually encounters the problems on how to set the initial condition and how to use the result obtained to prune a neural network. In this paper, some cues on the setting of the initial condition will be presented with a simple example illustrated. Then based on three assumptions--1) the size of training set is large enough; 2) the training is able to converge; and 3) the trained network model is close to the actual one, an elegant equation linking the error sensitivity measure (the saliency) and the result obtained via extended Kalman filter is devised. The validity of the devised equation is then testified by a simulated example.

On the regularization of forgetting recursive least square.

In this paper, the regularization of employing the forgetting recursive least square (FRLS) training technique on feedforward neural networks is studied. We derive our result from the corresponding equations for the expected prediction error and the expected training error. By comparing these error equations with other equations obtained previously from the weight decay method, we have found that the FRLS technique has an effect which is identical to that of using the simple weight decay method. This new finding suggests that the FRLS technique is another on-line approach for the realization of the weight decay effect. Besides, we have shown that, under certain conditions, both the model complexity and the expected prediction error of the model being trained by the FRLS technique are better than the one trained by the standard RLS method.

The effect of the nasopharyngeal air cavity on x-ray interface doses.

We investigated the impact of air cavities in head and neck cancer patients treated by photon beams based on clinical set-ups. The phantom for investigation was constructed with a cubic air cavity of 4 x 4 x 4 cm3 located at the centre of a 30 x 30 x 16 cm3 solid water slab. The cavity cube was used to resemble an extreme case for the nasal cavity. Apart from measuring the dose profiles and central axis percentage depth dose distribution, the dose values in 0.25 x 0.25 x 0.25 cm3 voxels at regions around the air cavity were obtained by Monte Carlo simulations. A mean dose value was taken over the voxels of interest at each depth for evaluation. Single-field results were added to study parallel opposed field effects. For 10 x 10 cm2 parallel opposed fields at 4, 6 and 8 MV, the mean dose at regions near the lateral interfaces of the cavity cube were decreased by 1 to 2% due to the lack of lateral scatter, while the mean dose near the proximal and distal interfaces was increased by 2 to 4% due to the greater transmission through air. Secondary build-up effects at points immediately beyond the air cavity cube are negligible using field sizes greater than 4 x 4 cm2. For most head and neck treatment, the field sizes are usually 6 x 6 cm2 or greater, and most cavity volumes are smaller than our chosen dimensions. Therefore, the influence of closed air cavities on photon interface doses is not significant in clinical treatment set-ups.

The effects of air cavities on X-ray dose distribution at 6 and 25 MV.

Three different air channels were constructed in solid water. These resemble upper respiratory and larynx geometry. Central-axis depth-dose distributions were measured on and beyond the distal surface of the air cavities. The measurements were made with a parallel-plate Markus ionisation chamber and TLDs for 6 and 25 MV X-ray beams of field sizes 4 cm x 4 cm and 4 cm x 7 cm. Measured dose was then compared to calculated dose for the Clarkson Scatter Integration and Equivalent Tissue-air Ratio (ETAR) algorithms. Both algorithms show errors in dose calculation at the distal surface of air cavities mainly because they fail to account for the effect of electronic disequilibrium. The magnitude of prediction error is found to depend on energy, field size and cavity geometry. For a 4 cm x 4 cm field at 6 MV the dose difference between the calculated value and the measured value, at the distal surface of square cavity is 8.2% and 13.7% for the ETAR and Clarkson algorithm respectively while that for a 4 cm x 7 cm field, the dose difference is reduced to 1.4% and 7.1% respectively. In general the ETAR algorithm shows better performance than the Clarkson algorithm because the Clarkson method calculates scatter dose assuming a homogeneous water medium while the ETAR uses a weighted sum of scatter components which is density dependent. The measured data can be used as benchmark data in the development and testing of new photon dose calculation algorithms.