[en] The inaugural use of CT and MRI imaging in brachytherapy, which traditionally has been limited to x-ray films, came during the development of a sophisticated implantation technique for brain cancer. Image-based techniques for prostate and gynecological implants have followed in the past decade. The same cardinal virtues of three dimensional treatment planning apply to image-based brachytherapy planning. Localization of structures in 3-D permits dose distributions to be indexed to individual anatomy, allowing conformal planning for the implant. Imaging during the implant produces precise execution of the plan, and post-implant imaging allows meaningful dosimetry that may be evaluated in relation to clinical outcome. Optimization of brachytherapy implant techniques may result, leading to enhanced clinical efficacy. The motivation for use and the application of 3-D image-based techniques for brachytherapy, with the emphasis on GYN intracavitary applications and prostate transperineal implants, will be discussed in this talk

[en] Purpose: Radiation therapy in combination with cis-platinum chemotherapy is associated with ototoxicity due to destruction of cochlear hair cells. This is a significant problem, especially in pediatric patients, because it may lead to difficulties with communication, speech, language, and development of learning skills. The use of 3D conformal radiotherapy (3D CRT) may be useful in sparing auditory structures. This paper discusses a technique using 3D CRT to spare the cochlea in patients with medulloblastoma. Methods and Materials: Five pediatric patients with medulloblastoma were planned using 3D CRT. All had MRI and CT obtained specifically for treatment planning. Multiple structures were contoured, including the cochlea and posterior fossa, and conformal beams designed in beam's eye view and dose distribution analysis were edited to provide 3D dose coverage to the target while sparing the inner ear. Patients received 36 Gy to the craniospinal axis followed by an 18-20 Gy boost to the posterior fossa. Results: A 3D CRT cochlear sparing technique was designed, using an axial pair of posterior oblique fields to treat the posterior fossa while sparing the cochlea for all patients in this analysis. Dose-volume information, obtained from 3D calculations, demonstrates that the average dose received by the cochlea was 65% of the prescribed dose using the cochlear sparing plan, as compared to 101% using standard opposed-lateral beams. Both plans delivered ≥ 100% of the prescribed dose to the posterior fossa. Conclusion: 3D CRT allows for cochlear sparing in the treatment of medulloblastoma. Further follow-up is necessary to determine the long-term benefit in these patients

[en] Purpose: Although the elective irradiation of regional nodal groups has gained widespread acceptance in the radiotherapy (RT) community as part of the standard treatment of NSCLC, elective nodal irradiation (ENI) is of unproven benefit and might actually compromise current efforts to improve the results of non-surgical treatment through intensification of RT regimens (i.e. dose escalation or altered fractionation regimens). By analyzing the computerized treatment planning data available for 10 patients treated on the University of Michigan lung dose escalation study, the current study quantifies 1) the increase in the irradiated volume, the effective volume of lung irradiated(V_eff), and the normal tissue complication probability (NTCP) of definitive RT when nodal groups are electively included in the clinical target volume, and 2) the extent to which inclusion of these elective nodal regions limits the ability to escalate RT dose. Methods and Materials: 72 patients have been treated on the phase I dose escalation study for NSCLC at the University of Michigan. All patients underwent CT scanning to obtain full volumetric image data. Therapy was planned on our three-dimensional radiotherapy treatment planning system. Once the optimal beam arrangement was determined, patients were assigned to one of five dose bins based on the lung effective volume (V_eff) of their treatment. Retrospectively, patients within each dose bin were ranked with respect to their V_eff and two patients who represented the median V_eff within each dose bin were selected for further analysis. For each patient selected, the V_eff and NTCP for pneumonitis were calculated for the following clinical target volumes (CTV): 1) gross tumor (primary plus gross nodal disease)(CTV-1); 2) gross tumor + mediastinum (CTV-2); 3) gross tumor + mediastinum + ipsilateral hilum (CTV-3); and 4) gross tumor + mediastinum + bilateral hila (CTV-4). All CTVs were expanded by 1 cm in three dimensions to create the planning target volume (PTV). V_eff was calculated for treatment of the elective nodal regions to 45 Gy using computer-planned APPA fields followed by a boost to PTV-1 to the dose the patient received on the protocol (as specified for his/her dose bin). All treatment plans were designed to encompass the PTV within the 95% isodose line(vol95). This volume encompassed by the 95% isodose lines was determined to reflect the volume of tissue irradiated for each PTV. Finally, to quantitate the extent to which ENI would compromise dose escalation, we determined the total dose the boost volume (PTV-1) could receive (following treatment of elective nodal regions to 45 Gy) in order to maintain an NTCP identical to the NTCP calculated when no elective nodal treatment was given.. Results: As shown in the table below, ENI (PTV 2, 3, 4) resulted in significant increases in the vol95, the V_eff, and the NTCP compared to treatment of the gross disease only (PTV-1). The data is presented as the mean ± S.E. Elective irradiation of the mediastinum (PTV-2) was associated with the greatest increase in vol95, V_eff, and NTCP. Addition of the ipsilateral hilum (PTV-3) and bilateral hila (PTV-4) to PTV-2 was associated with much smaller increases in these parameters. Analysis of the degree to which ENI compromises the ability to deliver high doses of radiation was also undertaken. It was clear that ENI resulted in a substantial limitation in the dose to which gross disease could be boosted -- in order to maintain the same probability for pneumonitis as that seen for irradiation of gross disease only (i.e. PTV-1), a 14% reduction in dose was required when the mediastinum was electively irradiated(i.e. PTV-2). As expected, the greatest compromise in dose was seen in the patients treated in the highest dose bins. Conclusion: By increasing the volume of lung irradiated, the V_eff, and the NTCP of treatment compared to irradiation of the gross tumor volume only, irradiation of elective nodal regions (especially the mediastinum) hinders our ability to intensify radiati on regimens for the treatment of NSCLC. Given our inability to control gross disease with current non-surgical treatment as well as the absence of data to support the benefit of ENI, the utility of elective irradiation of regional nodal groups to eradicate microscopic disease is questionable; thus, the need for ENI should be carefully considered prior to incorporation into protocols aimed at intensifying current non-surgical regimens

[en] Purpose: It is well known that the predominate pattern of failure of high grade astrocytomas is local recurrence. Using 3-dimensional conformal radiotherapy (3DCRT) high dose radiation can be delivered to a more precisely defined target while sparing normal tissue. However, if smaller target volumes are used to reduce morbidity, the risk for marginal misses may increase. The purpose of this study is to analyze the patterns of failure of high grade astrocytomas following high dose 3DCRT using a novel quantitative technique to calculate the dose received by the radiographically defined recurrence. Materials and Methods: From (4(89)) to (10(95)), 71 patients with supratentorial high grade astrocytomas have been entered in a dose escalation study. All patients were treated using 3DCRT to 70 - 80 Gy in conventional daily fractionation of 1.8 - 2.0 Gy. The clinical target volumes (CTV) consisted of successive cone downs with the final CTV defined as the enhancing lesion plus 0.5 cm margin. As of (10(95)), 45 patients have radiographic evidence of disease recurrence/progression. This is defined as 25% increase in the sum of products of measurable lesion over the smallest sum observed, reappearance of any lesion which had previously disappeared, or appearance of any new lesion. Of the 45 patients, 28 have recurrent scans (CT or MRI) that can be entered into our planning system and registered onto the treatment planning scans. Once the recurrent tumors were defined, dose volume histograms (DVHs) of the recurrent tumors were generated so that the dose delivered to the recurrent tumor volume from previous irradiation could be calculated. The recurrences were divided into 3 categories: 1) in-field recurrence, if ≥95% of the recurrence volume received ≥95% of the final prescribed dose, 2) marginal recurrence, if 26% - 94% of the recurrence volume received ≥95% of the final prescribed dose, and 3) distant recurrence, if ≤25% of the recurrence volume received ≥95% of the final prescribed dose. Results: The median age of the 28 patients was 51.5 years (range: 22.5 - 72.8 years). All had grade 4 supratentorial astrocytomas. The median time of disease recurrence/progression of these 28 patients was 4.1 months (range: 1.1 - 15.4 months). In 24 of 28 patients only one lesion was seen on recurrent scan. The remaining 4 patients had more than one lesion seen on recurrent scans. Of the 24 patients who had a single recurrence,(22(24)) were in-field recurrence and (2(24)) were marginal recurrences. The volumes of marginal recurrences that received ≥76 Gy (i.e., ≥95% of the prescribed dose) were 80.5% and 74%, respectively. In the 4 patients who had multiple recurrences, 3 patients had 2 lesions and 1 patient had 3 lesions. Their patterns of failure were as follows: in the first patient, one lesion was an in-field recurrence, but the second lesion was a marginal recurrence with 86% of the volume having received ≥76 Gy; in the second patient, both lesions were marginal recurrences with 41 - 77% of their volumes having received ≥76 Gy; in the third patient, both lesions were again marginal recurrences with 35 - 49% of their volumes having received ≥76 Gy, and in the last patient who had 3 recurrences, 1 lesion was a marginal recurrence with 69% of its volume having received ≥76 Gy but 2 lesions were distant recurrences with 0% of their volumes having received ≥76 Gy. Conclusion: This represents the first report which analyzed the patterns of failure of high grade astrocytomas with respect to the actual dose delivered to the volume occupied by the radiographic recurrence. In the current study we found that (23(28)) patients failed within the region of previous high dose irradiation (≥76 Gy). Of the five patients who had marginal recurrence as a component of their failure pattern, the volumes of the marginal recurrences that received ≥95% of the prescribed dose ranged from 35 - 86% with majority being >70%. No patients had distant recurrence alone. Therefore, despite high dose 3DCRT and the small margin applied to the contrast enhancing lesion for the final cone down, the predominate pattern remains one of local recurrence. Until the local disease can be controlled, further dose escalation using 3DCRT is warranted

[en] Purpose: Interstitial perineal implants may be utilized to deliver a high local radiation dose in the treatment of advanced gynecologic malignancies. Lack of knowledge of the precise anatomic relationships between the implant and the target and critical organs may limit efficacy and increase complication risks. Computed tomography (CT)-based planning, dose evaluation, and optimization of customized interstitial implants, aided by laparotomy, have been developed to overcome these limitations. Methods and Materials: Twenty patients with locally advanced gynecologic malignancies treated between May 1990 to October 1996 with external irradiation and one or two implants. Interstitial implants were performed when intracavitary brachytherapy was judged to be inadequate or when the response to external radiation and an intracavitary implant was not satisfactory. Customized interstitial implants were planned using preimplantation CT to determine catheter angles and paths that best implanted the target while avoiding pelvic bones and organs. Laparotomy aimed at lysing bowel adhesions, placement of omental carpet, and refining needle placement. Postimplantation CT was used for loading optimization and dose evaluation. Results: Catheter angles 15-25 deg. were found to adequately implant anteriorly laying targets while avoiding pubic bones and bladder. Adhesiolysis of bowel loops from the vaginal apex was required in patients with prior hysterectomy. Small modifications in catheter placements were made during laparotomy in all implants. Postimplantation CTs showed deviations of the catheter positions compared with the planning CTs and were essential in determining target and organ doses and loading optimization. At a median follow-up of 42 months (range: 9-80 months), local control rate is 55% and disease-free survival 40%. Late complications occurred in 2 of 11 of patients without local recurrence. Conclusions: CT-based planning, loading optimization, and dose evaluation of customized implants improve radiation dose delivery. Laparotomy enhances implant accuracy and safety. Local tumor control rate is still unsatisfactory. It reflects the shortcomings of technical advances alone in poor prognosis tumors like those selected for this series

[en] Purpose: To analyze the patterns of local-regional recurrence in patients with head and neck cancer treated with parotid-sparing conformal and segmental intensity-modulated radiotherapy (IMRT). Methods and Materials: Fifty-eight patients with head and neck cancer were treated with bilateral neck radiation (RT) using conformal or segmental IMRT techniques, while sparing a substantial portion of one parotid gland. The targets for CT-based RT planning included the gross tumor volume (GTV) (primary tumor and lymph node metastases) and the clinical target volume (CTV) (postoperative tumor bed, expansions of the GTVs and lymph node groups at risk of subclinical disease). Lymph node targets at risk of subclinical disease included the bilateral jugulodigastric and lower jugular lymph nodes, bilateral retropharyngeal lymph nodes at risk, and high jugular nodes at the base of skull in the side of the neck at highest risk (containing clinical neck metastases and/or ipsilateral to the primary tumor). The CTVs were expanded by 5 mm to yield planning target volumes (PTVs). Planning goals included coverage of all PTVs (with a minimum of 95% of the prescribed dose) and sparing of a substantial portion of the parotid gland in the side of the neck at less risk. The median RT doses to the gross tumor, the operative bed, and the subclinical disease PTVs were 70.4 Gy, 61.2 Gy, and 50.4 Gy respectively. All recurrences were defined on CT scans obtained at the time of recurrence, transferred to the pretreatment CT dataset used for RT planning, and analyzed using dose-volume histograms. The recurrences were classified as 1) 'in-field', in which 95% or more of the recurrence volume (V_recur) was within the 95% isodose; 2) 'marginal', in which 20% to 95% of V_recur was within the 95% isodose; or 3) 'outside', in which less than 20% of V_recur was within the 95% isodose. Results: With a median follow-up of 27 months (range 6 to 60 months), 10 regional recurrences, 5 local recurrences (including one noninvasive recurrence) and 1 stomal recurrence were seen in 12 patients, for a 2-year actuarial local-regional control rate of 79% (95% confidence interval 68-90%). Ten patients (80%) relapsed in-field (in areas of previous gross tumor in nine patients), and two patients developed marginal recurrences in the side of the neck at highest risk (one in the high retropharyngeal nodes/base of skull and one in the submandibular nodes). Four regional recurrences extended superior to the jugulodigastric node, in the high jugular and retropharyngeal nodes near the base of skull of the side of the neck at highest risk. Three of these were in-field, in areas that had received the dose intended for subclinical disease. No recurrences were seen in the nodes superior to the jugulodigastric nodes in the side of the neck at less risk, where RT was partially spared. Conclusions: The majority of local-regional recurrences after conformal and segmental IMRT were 'in-field', in areas judged to be at high risk at the time of RT planning, including the GTV, the operative bed, and the first echelon nodes. These findings motivate studies of dose escalation to the highest risk regions

[en] Purpose: Conformal treatment using static multisegmental intensity modulation was developed for patients requiring comprehensive irradiation for head and neck cancer. The major aim is sparing major salivary gland function while adequately treating the targets. To assess the adequacy of the conformal plans regarding target coverage and dose homogeneity, they were compared with standard irradiation plans. Methods and Materials: Fifteen patients with stage III/IV head and neck cancer requiring comprehensive, bilateral neck irradiation participated in this study. CT-based treatment plans included five to six nonopposed fields, each having two to four in-field segments. Fields and segments were devised using beam's eye views of the planning target volumes (PTVs), noninvolved organs, and isodose surfaces, to achieve homogeneous dose distribution that encompassed the targets and spared major salivary gland tissue. For comparison, standard three-field radiation plans were devised retrospectively for each patient, with the same CT-derived targets used for the clinical (conformal) plans. Saliva flow rates from each major salivary gland were measured before and periodically after treatment. Results: On average, the minimal dose to the primary PTVs in the conformal plans [95.2% of the prescribed dose, standard deviation (SD) 4%] was higher than in the standard plans (91%, SD 7%; p = 0.02), and target volumes receiving <95% or <90% of the prescribed dose were smaller in the conformal plans (p = 0.004 and 0.02, respectively). Similar advantages of the conformal plans compared to standard plans were found in ipsilateral jugular nodes PTV coverage. The reason for underdosing in the standard treatment plans was primarily failure of electron beams to fully encompass targets. No significant differences were found in contralateral jugular or posterior neck nodes coverage. The minimal dose to the retropharyngeal nodes was higher in the standard plans. However, all conformal plans achieved the planning goal of delivering 50 Gy to these nodes. In the conformal plans, the magnitude and volumes of high doses in noninvolved tissue were significantly reduced. The main reasons for hot spots in the standard plans (whose dose calculations included missing tissue compensators) were photon/electron match line inhomogeneities, which were avoided in the conformal plans. The mean doses to all the major salivary glands, notably the contralateral parotid (receiving on average 32% of the prescribed dose, SD 7%) were significantly lower in the conformal plans compared with standard radiation plans. The mean dose to the noninvolved oral cavity tended to be lower in the conformal plans (p = 0.07). One to 3 months after radiation, on average 60% (SD 49%) of the preradiation saliva flow rate was retained in the contralateral parotid glands and 10% (SD 16%) was retained in the submandibular/sublingual glands. Conclusions: Planning and delivery of comprehensive irradiation for head and neck cancer using static, multisegmental intensity modulation are feasible. Target coverage has not been compromised and dose distributions in noninvolved tissue are favorable compared with standard radiation. Substantial major salivary gland function can be retained

[en] Purpose: To analyze the failure patterns for patients with high-grade astrocytomas treated with high-dose conformal radiotherapy (CRT) using a quantitative technique to calculate the dose received by the CT- or MR-defined recurrence volume and to assess whether the final target volume margin used in the present dose escalation study requires redefinition before further escalation. Methods and Materials: Between 4/89 and 10/95, 71 patients with high-grade supratentorial astrocytomas were entered in a phase I/II dose escalation study using 3-D treatment planning and conformal radiotherapy. All patients were treated to either 70 or 80 Gy in conventional daily fractions of 1.8-2.0 Gy. The clinical and planning target volumes (CTV, PTV) consisted of successively smaller volumes with the final PTV defined as the enhancing lesion plus 0.5 cm margin. As of 10/95, 47 patients have CT or MR evidence of disease recurrence/progression. Of the 47 patients, 36 scans obtained at the time of recurrence were entered into the 3-D radiation therapy treatment planning system. After definition of the recurrent tumor volumes, the recurrence scan dataset was registered with the pretreatment CT dataset so that the actual dose received by the recurrent tumor volumes during treatment could be accurately calculated and then analyzed dosimetrically using dose-volume histograms. Recurrences were divided into several categories: 1) 'central', in which 95% or more of the recurrent tumor volume (V_recur) was within D₉₅, the region treated to high dose (95% of the prescription dose); 2) ''in-field,'' in which 80% or more of V_recur was within the D₉₅ isodose surface; 3) ''marginal,'' when between 20 and 80% of V_recur was inside the D₉₅ surface; 4) 'outside', in which less than 20% of V_recur was inside the D₉₅ surface. Results: In 29 of 36 patients, a solitary lesion was seen on recurrence scans. Of the 29 solitary recurrences, 26 were central, 3 were marginal, and none were outside. Multiple recurrent lesions were seen in seven patients: three patients had multiple central and/or in-field lesions only, three patients had central and/or in-field lesions with additional small marginal or outside lesions, and one patent had 6 outside and one central lesion. Since total recurrence volume was used in the final analysis, 6 of the 7 patients with multiple recurrent lesions were classified into central/in-field category. Conclusion: Analysis of the 36 evaluable patients has shown that 32 of 36 patients (89%) failed with central or in-field recurrences, 3/36 (8%) had a significant marginal component to the recurrence, whereas only 1/36 (3%) could be clearly labeled as failing mainly outside the high-dose region. Seven patients had multiple recurrences, but only 1 of 7 had large-volume recurrences outside the high-dose region. This study shows that the great majority of patient recurrences that occur after high-dose (70 or 80 Gy) conformal irradiation are centrally located: only 1/36 patients (with 7 recurrent lesions) had more than 50% of the recurrence volume outside the region previously treated to high dose. Further dose escalation to 90 Gy (and beyond) thus seems reasonable, based on the same target volume definition criteria

[en] Purpose: To compare seven techniques for irradiation of the postmastectomy chest wall (CW) using normal tissue complication probability (NTCP) predictions for pneumonitis and ischemic heart disease and dose-volume histogram analyses for normal and target tissues. Methods and Materials: Plan comparisons were performed for 20 left-sided postmastectomy CW RT cases using target volumes based on clinical delineation of standard field borders. Seven common treatment techniques were planned for each case, using a prescription of 50 Gy in 25 fractions to the CW and internal mammary node (IMN) targets. NTCP model metrics were used to quantify the risks of pneumonitis and ischemic heart disease, supplemented by dose-volume metrics to assess the target coverage to the CW and IMNs, as well as normal tissue dose (lung and heart). Results: Overlap in the distributions of the CW mean dose for all plans was found, except cobalt, which was significantly less than the remaining techniques (global F test, F=21.90, p<0.0001). Standard tangents produced a significantly lower IMN mean dose than all other methods, as expected (F=59.55, p<0.0001); the reverse hockey stick and cobalt techniques were lower than the other methods, which were statistically similar. Cobalt produced a significantly higher percentage of the heart that received >30 Gy (V30) than the other methods (F=49.76, p<0.0001). Use of partially wide tangent fields (PWTFs) resulted in the smallest heart V30. Use of cobalt fields resulted in a significantly greater NTCP estimate for ischemic heart disease than all the remaining techniques (F=70.39, p<0.0001). Standard tangents resulted in a percentage of the lung receiving >20 Gy (V20) significantly less than with PWTFs, 30/70 and 20/80 photon/electron mix, and reverse hockey stick techniques. NTCP estimates for pneumonitis revealed significantly better results with standard tangents (F=6.57, p<0.0001). Conclusion: No one technique studied combines the best CW and IMN coverage with minimal lung and heart complication probabilities. The choice of technique should be based on clinical discretion and the technical expertise available to implement these complex plans. Of the seven techniques studied, this analysis supports PWTFs as the most appropriate balance of target coverage and normal tissue sparing when irradiating the CW and IMN

[en] Multileaf collimator (MLC) systems are available on most commercial linear accelerators, and many of these MLC systems utilize a design with rounded leaf ends and linear motion of the leaves. In this kind of system, the agreement between the digital MLC position readouts and the light field or radiation field edges must be achieved with software, since the leaves do not move in a focused motion like that used for most collimator jaw systems. In this work we address a number of the calibration and quality assurance issues associated with the acceptance, commissioning, and routine clinical use of this type of MLC system. These issues are particularly important for MLCs used for various types of intensity modulated radiation therapy (IMRT) and small, conformal fields. For rounded leaf end MLCs, it is generally not possible to make both the light and radiation field edges agree with the digital readout, so differences between the two kinds of calibrations are illustrated in this work using one vendor's MLC system. It is increasingly critical that the MLC leaf calibration be very consistent with the radiation field edges, so in this work a methodology for performing accurate radiation field size calibration is discussed. A system external to the vendor's MLC control system is used to correct or handle limitations in the MLC control system. When such a system of corrections is utilized, it is found that the MLC radiation field size can be defined with an accuracy of approximately 0.3 mm, much more accurate than most vendor's specifications for MLC accuracy. Quality assurance testing for such a calibration correction system is also demonstrated