A. Dutreix

Combined radiotherapy and surgery: local control and complications in early carcinoma of the uterine cervix — the Villejuif experience, 1975–1984

From January 1975 to December 1984, 441 patients were treated by combined radiotherapy and surgery at the Institut Gustave Roussy (IGR) for Stage IB (288) and II (proximal) (103) carcinoma of the uterine cervix. Standard treatment consisted of pre-operative utero-vaginal brachytherapy (60 Gy) using a mould technique followed by a colpo-hysterectomy and external iliac lymphadenectomy. Overall 5 year actuarial survival for the whole population was 87% and disease-free survival 85%. Loco-regional relapse occurred in 23 patients (5%). Of these, 12 were central pelvic failures, 8 regional failures and 3 combined central and regional failures. There were 36 systemic relapses (8%) of which 12 relapsed concurrently in the pelvis. Five year actuarial pelvic disease-free, disease-free and overall survival was 87, 85 and 87%, respectively, for the whole population. 340 patients developed one or more complications [Grade 1: 198/441 (44%), Grade 2: 121/441 (27%) and Grade 3 or 4: 21/441 (4.7%)]. Five year actuarial survival for the whole population was poorer for histologically node positive than for node negative (89 vs. 55%, p less than 0.0001). Pre-operative brachytherapy followed by surgery can provide good local control with acceptable morbidity in early cervical cancer.

An approach to the interpretation of clinical data on the tumour control probability-dose relationship

A conventional statistical model allows predicting the sterilisation rate as a function of dose. However, the computation requires data on biological parameters (proportion of clonogenic cells, survival per fraction, multiplication rate) which are inaccessible for human tumours. The curative dose 50% (TCD50) can be used as a synthesis of these parameters and its significance for the response-dose relationship of a population of tumours of uniform radiosensitivity is discussed. The slope of the dose-control curve provides vital information regarding the variation in radiocurability of the various individual tumours. The model allows the analysis of the clinical data and the separation of tumour subsets with different radiation responsiveness. It provides an evaluation of the benefit which could be obtained from the identification of the subsets and a guidance for the clinical, pathological and biological studies which relate to this identification. The change in the response-dose relationships with the tumour size cannot (usually) be explained by the cell number increase alone. Other possible factors of reduced radiocurability are discussed.

A European quality assurance network for radiotherapy: dose measurement procedure

In the frame of the experimental implementation of a European quality assurance network for external radiotherapy, the methodology in the European Measuring Centre (MC) is presented. Mailed TL dosimeters are used for the check of the beam output and of the beam quality of photon beams. The thermoluminescent material is PTL 717 LiF powder. The readings were first performed on a manual, and then on an automatic reader, with standard deviations of the mean of 0.7% for one dosimeter. Corrections for supralinearity and for the energy dependence of the dosimeter response are applied. An original method has been developed to correct for the variation of the LiF response as a function of time. It is shown that the sensitivity of the powder changes during storage, leading to a kind of inverse fading. The global uncertainty of the TL postal measurement procedure is estimated to be about 1.5% for the 60Co beams and 2% for the x-ray beams. Intercomparisons with the IAEA and with the EORTC have shown an agreement better than 2% for all energies. It can be concluded that the results of the MC are suitable for the requirements of a European quality assurance network.

Can we compare systems for interstitial therapy

Three main dosimetry systems are used for interstitial brachytherapy: Manchester, Quimby and Paris Systems. A comparison is made of the rules of source distributions recommended by the three systems stating the advantages and disadvantages of each system. The comparison shows up the differences in the size of the treated volumes for similar implanted volumes. The next sections emphasize the differences in dose specification which make very difficult any comparison of clinical results obtained with different systems. Furthermore, the recent development of computers has lead some radiotherapists to specify the dose individually for each patient without reference to any published system. Very large variations in the dose delivered to a patient for a given implant and for the same prescribed dose, may results from this subjective procedure. In the last section one considers the various quality criteria proposed in the literature.

Keynote address: Prescription, precision, and decision in treatment planning

Abstract The goal of radiotherapy is to eradicate a tumor without causing severe damage to healthy tissues. Various published experiences have lead to the conclusion that an overall precision of ±5% on the absorbed doses, at any point in the patient, is required to meet this goal. Clear definitions of the method for specifying the absorbed dose and dose homogeneity throughout the target volume are essential, to facilitate communication, to improve the knowledge of dose-effect relations and to establish the necessary criteria for the optimization of treatment plans. Determination of the optimal energy is one of the most controversial problems in treatment plan optimization. It is clearly related to the criteria selected. A brief review of some criteria is proposed according to the tumor site. Computers may provide three-dimensional dose calculations for treatment conditions for which manual calculations are not feasible. Because the random errors are very small, computer calculations are often considered as exact although large risks of error are associated with each step of the calculation. The reduction of the overall uncertainty to the stated level of ±5% requires a constant effort from both radiotherapists and physicists at each step of treatment planning from basic definitions to dose distribution calculations.

The ESTRO-EQUAL Quality Assurance Network for Photon and Electron Radiotherapy Beams in Germany

Background: In 1998 an ESTRO Quality Assurance Network for radiotherapy (EQUAL) has been set up for 25 European countries for photon and electron beams in reference and non-reference conditions.nn Material and Methods: Measurements are done using LiF powder (DTL937-Philitech, France) that is processed with the PCL3 automatic reader (Fimel-PTW). The participating centers irradiate the TLDs with an absorbed dose of 2 Gy according to the clinical routine.nn Results: Until September 2000 EQUAL has checked 135 photon beams (including the beams rechecked) from 51 radiotherapy centers in Germany out of 86 accepted centers. The results show that 2% of the beam outputs in reference conditions and 3% of the percentage depth doses are outside the tolerance level (deviation > ± 5%). 6% of the beam output variations and of the wedge transmission factors show deviations > ± 5%. The global analysis of results shows deviations > ± 5% in at least one parameter for 18 beams out of the 135 beams checked. Five rechecked beams present one “real dosimetric” problem in one or more parameters, corresponding to 4% of the 114 beams for which the deviations cannot be attributed to set-up errors. – The EQUAL network has checked 89 electron beams in Germany. The results show that all beam outputs checked are within the tolerance level. The standard deviation for the beam output in reference conditions is 2.0% and 2.2% for the beam output for the others field sizes. The percentage of deviations > 3% and ≤ 5% for the reference beam output is higher for electron beams than for photon beam checks. Therefore the electron beam calibration and the TPS algorithms should be improved to increase the accuracy of the patient dosimetry for radiotherapy.nn Conclusion: EQUAL program demonstrates a consistency in radiotherapy dosimetry for photon and electron beams resulting in a satisfying accuracy of the dosimetry in Germany.Hintergrund: 1998 wurde für 25 europäische Länder ein ESTRO Quality Assurance Network (EQUAL) für die Strahlentherapie für Photonen- und Elektronenstrahlung unter Referenzbedingungen und davon abweichenden Bedingungen eingerichtet.nn Material und Methoden: Die Messungen erfolgen mit Lithiumfluoridpulver LiF DTL937 (Philitech, Frankreich), das mit der automatischen Auswerteeinheit PCL3 (Fimel-PTW) ausgewertet wird. Die teilnehmenden Zentren bestrahlen die TLDs der klinischen Routine entsprechend mit einer Energiedosis von 2 Gy.nn Ergebnisse: Bis September 2000 haben sich 86 Strahlentherapiezentren aus Deutschland beworben. Davon hat EQUAL 135 Photonenfelder (einschließlich der widerholt gecheckten Felder) von 51 Zentren geprüft. Die Ergebnisse zeigen, dass 2% der Dosiswerte unter Referenzbedingungen und 3% der prozentualen Tiefendosisdaten außerhalb des Toleranzniveaus (Abweichung > ± 5%) lagen, ebenso 6% der Daten bei Änderung der Dosiswerte mit der Feldgröße sowie des Keilfaktors. Die Analyse ergibt bei 18 der 135 geprüften Photonenfelder für mindestens einen Parameter Abweichungen > ± 5%. Fünf der erneut geprüften Felder zeigten ein “reales dosimetrisches” Problem bei einem oder mehreren Parametern, d. h., bei 4% von 114 Feldern können die Abweichungen nicht einem Fehler beim Messaufbau zugeordnet werden. – EQUAL hat in Deutschland 89 Elektronenfelder geprüft. Die Dosiswerte aller Felder lagen innerhalb des Toleranzbereichs. Die Standardabweichung für die Dosiswerte unter Referenzbedingungen betrug 2,0% und 2,2% für die Dosiswerte bei anderen Feldgrößen. Der Prozentsatz der Abweichungen > 3% und ≤ 5% für die Dosiswerte unter Referenzbedingungen war bei Elektronen größer als bei Photonen. Deshalb sollten die Kalibrierung der Elektronenfelder und die TPS-Algorithmen verbessert werden, um eine höhere Genauigkeit der Patientendosimetrie in der Strahlentherapie zu erreichen.nn Schlussfolgerung: Das EQUAL-Programm demonstriert eine Konsistenz in der strahlentherapeutischen Dosimetrie für Photonen- und Elektronenfelder mit einer befriedigenden Genauigkeit der Dosimetrie in Deutschland.

Energy correction factors of LiF powder TLDs irradiated in high-energy electron beams and applied to mailed dosimetry for quality assurance networks

Absorbed dose determination with thermoluminescent dosimeters (TLDs) generally relies on calibration in 60Co gamma-ray reference beams. The energy correction factor fCo(E) for electron beams takes into account the difference between the response of the TLD in the beam of energy E and in the 60Co gamma-ray beam. In this work, fCo(E) was evaluated for an LiF powder irradiated in electron beams of 6 to 20 MeV (Varian 2300C/D) and 10 to 50 MeV (Racetrack MM50), and its variation with electron energy, TLD size and nature of the surrounding medium was also studied for LiF powder. The results have been applied to the ESTRO-EQUAL mailed dosimetry quality assurance network. Monte Carlo calculations (EGS4, PENELOPE) and experiments have been performed for the LiF powder (rho = 1.4 g cm3) (DTL937, Philitech, France), read on a home made reader and a PCL3 automatic reader (Fimel, France). The TLDs were calibrated using Fricke dosimetry and compared with three ionization chambers (NE2571, NACP02, ROOS). The combined uncertainties in the experimental fCo(E) factors determined in this work are less than about 0.4% (1 SD), which is appreciably smaller than the uncertainties up to 1.4% (1 SD) reported for other calculated values in the literature. Concerning the Varian 2300C/D beams, the measured fCo(E) values decrease from 1.065 to 1.049 +/- 0.004 (1 SD) when the energy at depth in water increases from 2.6 to 14.1 MeV; the agreement with Monte Carlo calculations is better than 0.5%. For the Racetrack MM50 pulsed-scanned beams, the average experimental value of fCo(E) is 1.071 +/- 0.005 (1 SD) for a mean electron energy at depth Ez ranging from 4.3 to 36.3 MeV: fCo(E) is up to 2% higher for the MM50 beams than for the 2300C/D beams in the range of the tested energies. The energy correction factor for LiF powder (3 mm diameter and 15 mm length) varies with beam quality and type (pulsed or pulsed-scanning), cavity size and nature of the surrounding medium. The fCo(E) values obtained for the LiF powder (3 mm diameter and 15 mm length) irradiated in water, have been applied to the EQUAL external audit network, leading to a good agreement between stated and measured doses, with a mean value of 1.002 +/- 0.022 (1 SD), for 170 beam outputs checked (36 electron beam energies) in 13 reference radiotherapy centres in Europe. Such fCo(E) data improve the accuracy of the absorbed dose TLD determination in electron beams, justifying their use for quality control in radiotherapy.

Prescription, precision and decision in treatment planning

The goal of radiotherapy is to eradicate a tumor without causing severe damage to healthy tissues. Various published experiences have lead to the conclusion that an overall precision of +/- 5% on the absorbed doses, at any point in the patient, is required to meet this goal. Clear definitions of the method for specifying the absorbed dose and dose homogeneity throughout the target volume are essential, to facilitate communication, to improve the knowledge of dose-effect relations and to establish the necessary criteria for the optimization of treatment plans. Determination of the optimal energy is one of the most controversial problems in treatment plan optimization. It is clearly related to the criteria selected. A brief review of some criteria is proposed according to the tumor site. Computers may provide three-dimensional dose calculations for treatment conditions for which manual calculations are not feasible. Because the random errors are very small, computer calculations are often considered as exact although large risks of error are associated with each step of the calculation. The reduction of the overall uncertainty to the stated level of +/- 5% requires a constant effort from both radiotherapists and physicists at each step of treatment planning from basic definitions to dose distribution calculations.

On the use of a quality index to specify high energy photon beams

It is important to specify the beam quality in a simple and nonambiguous way in order on one hand to make comparisons easier between treatments performed in various hospitals, or at different times in the same hospital and on the other hand to facilitate the choice of numerical values for factors like restricted mass-collision stopping-power ratios and perturbation correction factors used in the conversion of ionization measurements into absorbed dose. We have adopted for high-energy photon beam specification a quality index (I) defined by the ratio (I20/I10) of ionizations measured with a constant source-detector distance for a reference field size 10 X 10 cm2. We have found that this quality index is independent of the source detector distance. On the other hand, the apparent linear attenuation coefficient measured on the exponential part of the tissue-maximum ratio curve can be calculated for any field size from the value of I for most high energy photon beams. In order to check the validity of the quality index for other linacs from other manufacturers, we have compared our results to published data related to various photon beams in a wide energy range: 2.5 to 45 MV.

Late effects of total body irradiation in correlation with physical parameters

Clear and complete documentation of the physical parameters of total body irradiation (TBI) is one of the essential requirements for the evaluation and improvement of the clinical results of TBI. Concerning the dosimetric aspects of TBI, a number of recommendations have been formulated with emphasis on basic dosimetry, patient dosimetry and dose specification. The dosimeters should be calibrated regularly with reference to the absorbed dose in water. Depth dose measurements should be performed in water equivalent phantoms of specified dimensions. It has been strongly suggested to measure the absorbed dose at the surface of the patient at 8 different regions at the entry and exit of the beam under TBI conditions. The reference dose to the patient should be specified as the total dose to mid abdomen at the height of the umbilicus. As an independent parameter, the lung dose should be specified as the mean dose in the central region of the shielded part of both lungs. Recent, more complete, information on the physical and dosimetric aspects of TBI will be incorporated in the registry of the European Bone Marrow Transplant Group (EBMT). A cooperation has been established between the EBMT and the European Late Effects Project Group (EULEP) to study the development of late effects in man caused by ionising radiation.