Frank M. Waterman

Response of human tumor blood flow to local hyperthermia

The effect of heat on blood flow in human tumors was studied as a function of time during 1 hour of local hyperthermia induced by 915 MHz microwaves. Blood flow was determined from the rate of thermal clearance by use of the bio-heat transfer equation. The rate of thermal clearance was measured at intervals of approximately 10 minutes throughout the treatment session by turning off the microwave power for 50 seconds. Tumor blood flow increased by amounts varying from 15 to 250% during the first 20-50 minutes of heating at 41-45 degrees C, after which it remained relatively constant during the remainder of the treatment session. The sharp reduction in blood flow or vascular stasis reported in most transplantable rodent tumors after comparable heating was not observed in human tumors. The maximum blood flow observed in heated human tumors ranged from 10-40 ml/min/100 gm. The systematic error due to thermal conduction was estimated to be equivalent to a blood flow of less than 3 ml/min/100 gm.

Limitations of the minimum peripheral dose as a parameter for dose specification in permanent 125I prostate implants

PURPOSEnThe objective of this work is to investigate whether the minimum peripheral dose is a practical parameter for dose specification in permanent 125I implants of the prostate.nnnMETHODS AND MATERIALSnThe investigation was carried out by use of a computer model of ellipsoidal 125I implants in which the average dimension and elongation factor were varied to provide a wide range of geometries. Both ideal and nonideal implants were investigated. The 125I seeds were confined to the target volume except for a portion of the study in which the effect of placing seeds outside the target volume was investigated.nnnRESULTSnThe minimum peripheral dose was found to be very sensitive to the seed placement. The irregularities in the seed spacing that inevitably occur in actual implants tend to lower the minimum peripheral dose. As a result, the minimum peripheral dose is generally significantly less than planned by an amount that is unpredictable, and often exceeds 25%. However, the percentage of the target volume that receives a dose less that the prescribed minimum peripheral dose is generally less than 10%. Implanting seeds outside the target volume improves the dose uniformity, but does not appear to offer any advantage in dose coverage, and increases the volume of normal tissue irradiated.nnnCONCLUSIONnIf a minimum peripheral dose is prescribed for a permanent 125I prostate implant, and the implant is planned using an idealized implant having precisely spaced seeds, the prescribed dose will rarely, if ever, be achieved. Reasonable agreement with the prescribed dose can be achieved only if the requirement for coverage is relaxed from 100 to 90%, or if the total source strength is increased by 20% to compensate for the anticipated imperfections in seed placement.

Cholangiocarcinoma: The Impact of Tumor Location and Treatment Strategy on Outcome

The purpose of this study was to evaluate how the outcome of patients with extrahepatic cholangiocarcinoma (EHBC) may have been influenced by tumor location and treatment selection. The primary endpoint of this study is overall survival (OS). Between January 1983 and December 1997, 221 patients with biliary tumors were evaluated at Thomas Jefferson University Hospital. Of these, 118 fit the inclusion criteria for this study. The extent of disease was assessed by computed tomography, percutaneous transhepatic cholangiography or endoscopic retrograde cholangiopancreatography, magnetic resonance imaging, and ultrasonography. All patients had histologic confirmation of malignancy. Roux-en Y, hepaticojejunostomy, or choledochojejunostomy followed surgical resection of the primary tumor. Palliative measure (PS) included biliary catheter placement without brachytherapy or external beam irradiation (RT). RT was delivered via high-energy photons. Intraluminal brachytherapy was performed via percutaneous biliary catheterization with iridium-192 ribbon sources. Chemotherapy consisted of either intravenous 5-fluorouracil alone or in combination with doxorubicin, mitomycin C, or paclitaxel. PS consisted of metal bile duct stent placement. Median follow-up time for the entire group was 102 months and 43 months for patients who were still alive at the conclusion of the study period. Patients with proximal tumors underwent resection (n = 5), surgery and RT (n = 23), RT only (n = 31), chemotherapy only (n = 6), or PS (n = 12). Patients with distal tumors were treated with surgical resection (n = 17) or a combination of surgery and RT (n = 13), RT only (n = 6), or PS (n = 4). Median survival time (MST) for all 118 patients was 22 months. The MST for patients with distal tumors was 47 months versus 17 months for those with proximal tumors. The MST has not been reached for patients with distal EHBC treated with surgical resection and postoperative RT, whereas the median survival for those treated with surgery alone is 62.5 months. However, 4 of 17 of these patients had in situ carcinoma. Six patients had distal tumors treated with RT only with a MST of 6 months. Patients with proximal tumors treated with surgery and RT had a superior OS at 5 years compared to patients treated with RT alone (24 vs. 13 months; p = 0.007). There was an improved OS in patients with proximal tumors treated with surgical resection and RT compared to surgery alone (p = 0.023). There is no discernable influence of chemotherapy on outcome in patients with proximal EHBC. The MST for patients treated with PS was 3.5 months. Surgery and postoperative RT appear to be better than either surgery or RT alone in patients with proximal EHBC. In patients with distal EHBC, the addition of resection and RT appears to offer an advantage, which is increasingly apparent with longer follow-up time. The prognosis remains dismal for patients treated with palliative intent.

Blood flow in human tumors during local hyperthermia.

The response of tumor blood flow during local hyperthermia was studied at 40 different points in 15 superficial human tumors. Hyperthermia was administered for 60 minutes by use of 915 MHz microwaves. Blood flow was determined from the rate of thermal clearance by use of the bioheat equation. The rate of thermal clearance was sampled at 10-15 minutes intervals by turning the applied power off for 30 seconds. A correction was made for thermal conduction from orthogonal profiles of the tumor temperature. No measurements were made during the first 10-15 minutes of heating. The response of tumor blood flow was found to be independent of temperature in the range of 40-44 degrees C. The mean blood flow rate increased 10-15% between 15 and 30 minutes, but remained nearly constant thereafter. The coefficient of variation in this pattern is 15-20%. No evidence of a sharp reduction in flow was observed. Furthermore, the mean temperature elevation, net forward power, and rate of thermal conduction all remained nearly constant with time, providing further evidence of stability in the blood flow rate. Data obtained in one tumor suggest that a reduction in flow may occur at temperatures above 44 degrees C. The mean blood flow rates obtained in this study range from 0-34 ml/100g/min with an average value of 15 ml/100g/min.

Determination of the temperature artifact during ultrasound hyperthermia

Temperature artifacts produced by very small uncoated thermocouples during ultrasonic heating are evaluated by backward extrapolation of the linear portion of the temperature rise curve or by backward extrapolation of the exponential portion of the temperature decay curve. The accuracy of these techniques for larger clinically used thermocouples is investigated by use of a two-dimensional model of the bioheat equation which simulates the transfer of heat radially from a probe 1 mm in diameter. The accuracy of these techniques is found to depend upon the perfusion rate. In the absence of perfusion, both extrapolation techniques underestimate the artifact by nearly 40%. Extrapolation of the temperature rise curve is very sensitive to the perfusion rate and this technique results in errors exceeding 100% when the perfusion rate is high. Extrapolation of the temperature decay curve produces more consistent results. Over a blood flow range of 0-100 ml/100 g per min, the artifact is underestimated by an amount that varies from approximately 40% to 30% respectively. Thus, the artifact can be determined to within 5% by this technique by increasing the extrapolated value by 35%.

Catheter induced temperature artifacts in ultrasound hyperthermia.

Temperature artifacts were evaluated at 72 different sensor locations in 10 different tumour sites heated by use of planar ultrasound transducers operated at 1 and 3 MHz. Thermometry was carried out by single- and multisensor thermocouple probes inserted into 19- and 16-gauge polyurethane catheters, respectively. Nearly all catheters were oriented approximately perpendicular to the ultrasound beam. The artifacts were determined by backward extrapolation of the thermal decay 30-60s after the power was turned off. The effective blood flow and specific absorption rate (SAR) at the sensor locations were determined from the rate of decay and the steady-state temperature. The sample mean steady-state temperature, effective blood flow, and SAR were 41.4 degrees C, 17.5 ml/100 g/min, and 46.3 W/kg, respectively. The most frequent artifact was in the range 0-0.2 degrees C and the mean artifact was 0.6 degrees C. Less than 15% of the artifacts were above 1 degree C. The magnitude of the artifact correlates with the SAR of ultrasonic power, the effective blood flow rate, and the steady-state temperature. These results indicate that the artifact produced at 1 MHz by a multisensor, Teflon-sheathed thermocouple inserted into a 16-gauge polyurethane catheter is 1.7 +/- 0.4 degrees at an SAR of 100 W/kg.

Mechanisms of heat removal during local hyperthermia

Mechanisms of heat removal were studied in five recurrent squamous cell head or neck carcinomas, 50-150 cm3, heated by use of external 915 MHz microwave applicators. Thermal clearance measurements were made at a single point in each tumor. Three profiles of the tissue temperature were also measured in orthogonal directions about this point. The conduction term of the bioheat equation was evaluated from the orthogonal temperature profiles by the method of finite differences. The perfusion term of the bioheat equation was determined from the rate of temperature decay corrected for conduction. The results show that thermal conduction plays a major role in the dissipation of thermal energy during local hyperthermia. The rate of removal thermal energy by conduction ranged between 20 and 150 percent of that by perfusion. The temperature profiles show that conduction is higher than is generally expected due to heterogeneities in the blood flow which produce rapid changes in the temperature gradient. The results of this study demonstrate that the heat transport by thermal conduction in perfused tissue cannot be assumed to be small, or negligible, in comparison to that by perfusion.

The effect of coupling materials on specific absorption rate distributions at 915 MHz.

Deionized water is commonly used to couple microwave applicators to the patient surface in the administration of local hyperthermia. Profiles of the specific absorption rate (SAR) at 1-cm depth show that deionized water coupling significantly distorts the SAR distributions of the Clini-Therm 915-MHz 10 X 10 and 15 X 15 cm2 applicators. Maxima and minima that are discernible in the SAR profiles obtained by direct applicator load contact are amplified producing unexpected hot and cold regions in the heating pattern. An exception is coupling achieved by use of the Clini-Therm cooling pad, oriented such that the direction of deionized water flow is perpendicular to the electric field. The distortion in the SAR distribution can also be eliminated by replacing deionized water with mineral oil, a material having a much lower dielectric constant (epsilon = 2). The SAR profiles for mineral oil coupling are comparable to those obtained for direct contact; however, the efficiency of power transfer is slightly less (70%) and the level of microwave leakage is approximately four times greater.

A recommended revision in the RTOG thermometry guidelines for hyperthermia administered by ultrasound

RTOG thermometry guidelines for clinical trials of hyperthermia using planar ultrasound recommended that temperatures be mapped in polyurethane catheters by use of single-junction copper-constantan thermocouples. These guidelines were based on an assumption that the error in temperature measurement due to thermal conduction would generally not exceed +/- 0.3 degrees C. The validity of this assumption was tested with a commercially available single-junction copper-constantan thermocouple. The width of the point spread function, an indicator of the relative magnitude of the conduction error, was five times greater than expected. As a result, the conduction error is projected to exceed 0.3 degrees C in a temperature gradient of only 1.5 degrees C/cm. This projection was confirmed by mapping a thermal peak which simulates a typical clinical temperature profile. This peak had an amplitude of 6 degrees C, a full-width at half-maximum of 3.5 cm, and a maximum gradient of approximately 3 degrees C/cm. Temperatures measured at 0.5-cm intervals over the span of this peak were in error by a mean of +/- 0.6 degrees C. It is strongly recommended that the RTOG guidelines be revised to replace copper-constantan thermocouples with manganin-constantan single- or multi-junction thermocouples which will assure that the conduction error will be < +/- 0.3 degrees C.

Use of manganin-constantan thermocouples in thermometry units designed for copper-constantan thermocouples

Commercial ultrasound hyperthermia systems typically include thermometry units designed for copper-constantan thermocouples. Replacing these copper-constantan thermocouples with manganin-constantan thermocouples is advantageous in reducing the measurement error caused by the conduction of heat along the copper wire, but their performance in these thermometry units is uncertain. The accuracy of manganin-constantan thermocouples in the Labthermics LT-100, Clini-Therm TS1200/TM100, and Physitemp TM-12 thermometry units was investigated using a temperature controlled circulating water bath monitored by a mercury thermometer having a calibration traceable to NIST. The results demonstrate that an accuracy of +/- 0.2 degrees C can be achieved with manganin-constantan thermocouples over the range 35-55 degrees C without hardware modification provided specific calibration procedures are followed. With the Labthermics LT-100, a double point calibration should be carried out at 35 and 55 degrees C. With the Clini-Therm TS1200/TM100, a self-calibration of the unit using its internal calibration well plus a single point calibration using an external temperature standard provides sufficient accuracy. The Physitemp TM-12 requires an external computer for read out and the user must provide additional software to correct for the error by either a single or multiple point calibration.