Advertisement

Robust optimization used to mitigate setup uncertainties in vulvar patients receiving VMAT

Published:December 31, 2021DOI:https://doi.org/10.1016/j.meddos.2021.11.003

      Abstract

      Many radiation planning techniques have been used to increase dose to the vulvar surface when treating patients with vulvar cancer with volumetric modulated arc therapy. Target volumes near the skin surface, such as vulvar tumors, do not meet the International Commission on Radiation Units and Measurements safety guidelines. Without the needed expansions and setup uncertainties, there are concerns that treated dose to the vulvar surface varies from the planned dose. The purpose of this study was to determine if the robust optimization tool used in volumetric modulated arc therapy planning can reproducibly deliver prescription dose to the vulvar skin surface despite setup uncertainties and reduced safety margins. To further mitigate the possibilities of setup errors and dose variation to the vulvar surface, robust optimization during planning must be considered. For this case study, 5 patients with similar diagnoses and vulvar surface dose requirements were retrospectively replanned using robust optimization. An evaluation structure was created representing the vulvar surface for each patient. This structure was assigned robustness characteristics. After multiple optimizations and planning scenarios, perturbed dose representing dataset shifts were evaluated along the vulvar surface. Every patient met minimal metrics of 100% of vulvar surface volume receiving 95% of the prescription dose. An additional review of 95% of vulvar surface volume receiving 100% of the prescription dose showed similar results. Robust optimization can be used in the planning process to mitigate setup uncertainties.

      Keywords

      Introduction

      Vulvar cancers are relatively rare, accounting for < 0.5% of new cancers diagnosed yearly and approximately 4% of all gynecological cancers. Vulvar cancers are squamous cell skin cancers most commonly involving the labia majora and/or minora making treatment options complex and customized.
      • Weinberg D
      • Gomez-Martinez RA.
      Vulvar cancer.
      Treatment options for vulvar cancer include surgery, chemotherapy, and radiation therapy.
      Using radiation therapy to treat vulvar cancer has posed challenges in treatment planning. Historically, gynecological cancers were treated with 3D conformal radiation therapy that included a large anterior photon field, a smaller posterior field and bilateral anterior electron fields which were used supplement dose to the anterior inguinal nodes. In addition, bolus was used to increase surface dose to the vulva and would commonly be removed when patients showed evidence of skin desquamation or erythema. Three-dimensional planning techniques provided adequate coverage of the tumor and respective lymph nodes (LNs); however, normal tissue toxicity was of concern. Advancements in radiation therapy, such as intensity modulated radiation therapy and volumetric modulated arc therapy (VMAT) have provided more efficient, conformal treatment planning. Modulated radiotherapy has become the standard in treating vulvar patients as the benefits include tumor conformality, dose escalation and organs at risk sparing.
      The International Commission on Radiation Units and Measurements (ICRU) has established target volume contouring guidelines to increase accuracy and standardization in radiation therapy planning. Fundamentally, the ICRU defined any visible disease as the gross tumor volume (GTV) and subsequently, a margin was added to the GTV to account for potentially microscopic disease and noted as the clinical target volume (CTV).
      Library of Congress Cataloging-in-Publication Data
      International Commission on Radiation Units and Measurements. ICRU Report 62.
      Both GTVs and CTVs were used to target disease; however, further margins were required to account for potential errors in target location. A Set-up Margin (SM) was subsequently defined and included variations in patient positioning, mechanical uncertainties of equipment, dosimetric uncertainties, transfer set-up errors from CT simulators to the treatment unit and human errors.
      Library of Congress Cataloging-in-Publication Data
      International Commission on Radiation Units and Measurements. ICRU Report 62.
      All 3 volumes were combined to form a planning target volume (PTV). Planning target volumes are determined by the physician in treatment planning and prescribed to ensure that the GTV and/or CTV will receive adequate dose coverage. Typical PTV margin expansions are not problematic unless the GTV and/or CTV are adjacent to patient's skin surface or external contour. In such cases, desired expansion is not possible as the expansion will extend beyond the skin or external surface. In the case with vulvar cancers the expansion is truncated at the skin surface. The inability to include standard uncertainties in the PTV contour present significant difficulties in treatment planning and treatment delivery.
      Though VMAT is the preferred planning technique for vulvar cancer patients, delivering dose to the skin surface without the adequate PTV margins accounting for setup uncertainties is a challenge. Conceptual planning techniques such as integrated skin flash (ISF) with virtual bolus have been used to deceive the optimizer and improve coverage at the skin surface lacking PTV margins. For the integrated skin flash technique, a 1.5 to 2.0 cm of virtual bolus was created.
      • Dyer BA
      • Jenshus A
      • Mayadev JS.
      Integrated skin flash planning technique for intensity-modulated radiation therapy for vulvar cancer prevents marginal misses and improves superficial dose coverage.
      In addition to the virtual bolus, a planning structure was also created. The planning structure was copied from the virtual bolus structure and its outer edge was retracted 5.0 mm from the outer edge of the virtual bolus. The virtual bolus and planning structure were used during optimization only to help mitigate and control higher dose gradient at the interface of the virtual bolus surface and air. Assigning appropriate constraints to the planning structure during optimization prevents the multileaf collimators from closing to the vulvar skin surface and ensures adequate dose when delivering modulated radiotherapy.
      Integrated skin flash (ISF) technique is analogous to skin flashing techniques used to plan breast cancer patients who received modulation radiotherapy. The ISF technique allows for intra- and inter-fraction motion when skin dose is a priority. However, the limitations of the virtual bolus technique are noticeable during the final calculation with the virtual bolus removed. When the virtual bolus is removed and replaced with a clinical bolus of 3.0 to 5.0 mm, the final calculation will show increased hotspots along the skin surface of the perineum and vulva. The hotspots displayed within the plan vary daily due to setup and are typical when using ISF planning techniques to achieve adequate skin dose. In addition, the treatment planning system (TPS) equates the absence of virtual bolus in the final calculation as missing tissue and may contribute to higher skin toxicity. Due to the possibility of higher skin toxicity in using the ISF and virtual bolus techniques, improved planning techniques must be considered.
      Robust optimization is a new technology that can be used during the planning process to increase dose and improve dose homogeneity at the vulvar skin surface without the use of ISF and virtual bolus. During the robust optimization process, the TPS applies user defined shifts in the x, y, and/or z direction as well as shifts in the CT density curve to account for setup uncertainties for the selected structure(s). The optimizer accounts for setup uncertainties while optimizing dose to the structure(s). After calculation, the TPS displays the uncertainties as multiple planning scenarios that represent the user defined shifts. In short, the robust optimization technique could apply and correct for target volume setup scenarios that do not have appropriate SMs, such as the case with vulvar patients. Robust optimization can decrease dosimetric variations observed across the vulvar skin surface while further increasing treatment planning efficiency for medical dosimetrists.
      As radiation therapy technology evolves, so does the planning and treatment delivery of vulvar cancer. Though target margins are standardized within the industry, the problem is that standardized PTV margins are designed to account for setup uncertainties which are not attainable along the skin surface for vulvar cancer. The ISF technique in combination with virtual bolus may be adequate as a skin flashing tool during optimization to achieve skin dose, yet the problem of skin toxicity remains. To date, there is a paucity of literature on alternative VMAT planning strategies, such as robust optimization, to address vulvar skin dose. The purpose of this study was to determine if the VMAT treatment planning robust optimization tool could deliver a reproducible, homogenous prescription dose to the vulvar skin surface despite setup uncertainties and reduced safety margins. The goals of this research were to determine if the robust optimization technique could (1) deliver ≥ 95% of the prescription dose to 100% of the skin surface in the least desirable setup scenario, (2) deliver ≥ 100% of the prescription dose to 95% of the skin surface in the least desirable scenario and (3) not deliver ≥ 110% of the prescription dose to 0.03 cubic centimeters (cc) at the skin surface.

      Methods and Materials

      Patient selection & setup

      Patient selections was made retrospectively from a small population of patients diagnosed with vulvar squamous cell cancer (SCC). Inclusion criteria were patients with unresectable SCC of the vulva, post radical vulvectomy with deep invasion and/or close margins, or recurrent vulvar disease, and prior VMAT to the pelvis where vulvar skin dose was a concern. Five patients between the ages of 34 and 78 years old with International Federation of Gynecology and Obstetrics (FIGO) staging IB and higher were chosen for this case study and were identified as patients A – E (Table 1). Patients were simulated in the supine position and immobilized in a frogged-legged position using a vac-lok bag. All 5 patients in this case study were simulated with a full bladder to minimize dose to the small bowel.
      Table 1Vulvar cancer patient prescription and fractionation in case study
      PatientStageInitialCD 1CD 2CD 3PTV_EVAL
      AClinical Stage III (pT2b,N2)45 Gy in 25 fxs50 Gy in 25 fxs54 Gy in 25 fxs2 Gy in 1 fxs10 Gy in 5 fxs6 Gy in 3 fxs

      50 Gy in 25 fxs
      Pelvis Nodes, VaginaVulva, bilat inguinalAll Gross NodesVulva (52 Gy)Bilat Gross Nodes (56 Gy)Vulva (62 Gy)Gross Nodes (66 Gy)Vulva (68 Gy)
      BFIGO IB (pT1b,pN0)45 Gy in 25 fxs50 Gy in 25 fxs10 Gy in 5 fxs

      50 Gy in 25 fxs
      Pelvic NodesVulva InguinalVulva
      CRecurrent, unresected SCC vulva54 Gy in 30 fxs51 Gy in 30 fxs60 Gy in 30 fxs

      60 Gy in 30 fxs
      Lt GroinBilat Pelvic NodesVulva RT Inguinal
      DRecurrent, unresected SCC vulva52.8 Gy in 33 fxs67.65 Gy in 33 fxs

      67.65 Gy in 33 fxs
      Vagina GTV +1.5 cm, paravaginal tissue to pelvic sidewalls, regional nodes (inguinal, external/internal iliac nodes, presacral)Vulva, Vagina GTV+1.0 cm
      EStage III invasive SCC (pT1b,pN1b(SN), M047.6Gy in 28fxs56Gy in 28fxs60.2 Gy in 28 fxs

      56 Gy in 28 fxs
      Pelvic lymph nodesVulva, clitoris and 3.0 cm left labia/muscleMultiple lymph nodes; bilateral groin and pelvis
      CD1, cone down 1; CD2, cone down 2; CD3, cone down 3; Gy, Gray (dose); FIGO, international federation of gynecology and obstetrics; Fxs, fractions; M, metastatic disease; N, lymph nodes involvement; P, pathologic stage; PTV_EVAL, evaluation PTV; T, tumor size; SCC, squamous cell carcinoma; SN, sentinel lymph node biopsy.

      Target delineation

      The radiation oncologist delineated tumor volumes. Structures were labeled based on anatomic area or prescription dose. In addition to active disease, prophylactic nodes were contoured and treated based upon risk factors and relapse probability. In all cases, the vulvar region, vagina, bilateral inguinal and regional LN were delineated. The regional LNs included para-aortic, common iliac, external, and internal iliac nodes. In more advanced cases, the presacral nodes were also included. Expansions varied per patient but averaged between 5.0 to 7.0 mm for LNs; 7.0 to 15.0 mm for gross or microscopic disease. Inguinal LN expansions were constricted 3.0 to 5.0 mm from the external skin surface (Fig. 1). Vulvar GTV or CTV expansions were 10.0 to 15.0 mm but truncated at the skin surface. Retrospectively, a PTV_EVAL_vulvar_skin (PTV_EVAL) structure was created from the vulvar target volume. The PTV_EVAL included the most anterior 5.0 mm of the vulvar TV, adjacent to skin (Fig. 1). The PTV_EVAL structure was created for use in the robust optimization and the evaluation of dose to the vulvar skin surface. Organs at risk were contoured by the medical dosimetrist which included bladder, rectum, sigmoid, small bowel, bowel bag, and bilateral femoral heads.
      Fig 1
      Fig. 1Computed tomography axial, sagittal and coronal planes representing the planning target volume (PTV) and bolus on 2 patients. A). Structures are labeled as PTV_5400 (red), PTV_5000 (green), PTV_4500 (blue), and GTV_Vulva (light red). B). Structures represent a newly created PTV_EVAL_Vulvar Skin (PTV_EVAL) structure (pink), clinical bolus (yellow), CTV_Vulva_5000 (cyan), and PTV_Vulva_5000 (orange). (Color version of figure is available online.)

      Treatment planning

      Original VMAT plans were completed in either Pinnacle 9.1 or Raystation9B.
      The planning techniques used in the original VMAT plans varied between dosimetrists. Some of these techniques included ISF optimization, prescribed dose to pseudo-structures, adding bolus at the time of simulation, or adding a bolus during the planning process. However, all patients were clinically treated with either partial or complete bolus over the vulvar surface area. Prescriptions and planning were unique and customized to each patient in this case study (Table 1). Fractionation regimens ranged from 25 to 34 fractions. Dose to the vulvar surface varied from 50 to 67.7 Gy depending on staging. Some patients received prescription dose with Simultaneous Integrated Boost- intensity modulated radiation therapy while other patients were treated with consecutive cone-down plans. Prospective plans using robust optimization were completed on every patient in Raystation9B.
      Robust optimization plans were generated on 5 patients to investigate dose distribution along the vulvar surface when setup uncertainties are a concern and could affect the dose delivered compared to the dose planned. Structurally, the PTV_EVAL structure was created to mimic an internal target volume (ITV), as seen in robust planning for lung patients. In other studies that used robust optimization, the ITV replaced the PTV allowing the optimizer the ability to correct for motion and setup errors simply by running a multitude of plans simultaneously representing possible tumor movement, changes in shape or size of a tumor and/or setup uncertainties.
      Computer-generated dataset shifts, the perturbed dose, and the robust optimization blurred dose at the vulvar surface was not reflective of the therapeutic dose coverage to the GTV or CTV that already had acceptable expansions (PTV) and/or SMs. The missing expansion anterior and beyond the skin surface were closely observed as a key structure during the optimization process. The PTV_EVAL was added to the optimization criteria. With the assignment of a robust factor, the PTV_EVAL was heavily weighted (proportionally higher than other target volumes and other non-robust structures) and requested minimum and uniform dose values at the prescription dose and a maximum dose of 105.0%. The medical dosimetrist defined intervals of deviation during optimization, creating an area of uncertainty and weighted factors to aide in delivering the planned dose. Using this technique, the TPS was able to account for shifts in the CT simulation dataset. Utilizing robust optimization eliminated the need to create pseudo or optimization structures (rings or avoidance structures), to force dose onto the skin surface.
      All TVs were evaluated and identified as having or not having adequate SMs. The structures that did not have acceptable expansion were allocated as robust optimization structures. All 5 patients were replanned using robustness optimization with a newly designed PTV_EVAL and assigned as a robust structure. The robust optimization settings were entered in the TPS. Patient position uncertainty was defined as the parameters of 1.0 cm isotropic. Robust optimization created 21 planned scenarios that balanced dose distribution with considerations for setup uncertainties. Selecting multiple GTVs and CTVs as robust structures for optimization would be unnecessary if ICRU guidelines could be followed in creating SMs such as PTV. Applying robustness to many structures unnecessarily would decrease efficiency and increase calculation time. Limiting the selection of which structure(s) need robustness is important in decreasing the time between the simulation and the patient's first treatment. Therefore, GTV and CTV requiring robustness should be limited to target volumes that do not have a geometric SM or PTV. Each patient was treated on the Elekta Versa linear accelerator using 6 MV energies, 2 full arcs, 0.2 × 0.2 × 0.2 cm grid and 3.0 to 5.0 mm slab bolus over the vulvar skin area during treatment.

      Planning evaluation

      The PTV_EVAL structure was evaluated using the robust optimization tool delivering ≥ 95% of the prescription dose to the skin surface in the least desirable setup scenario. Robust optimization can spread the dose distribution across an area where the tumor motion or patient's surface maybe in question. Plan evaluation was done by reviewing the dose distribution, dose volume histogram, and scorecards (dose metrics) for every patient in this study. Traditional VMAT plans, when compared to those using robust optimization, were very similar in comparison (Fig. 2). Robust optimization was performed to maintain plan integrity while delivering clinical patient treatments with intrinsic setup uncertainties. These uncertainties were evaluated by computing perturbed dose on the patient datasets after defined shifts in the x, y, and/or z direction were performed.
      Fig 2
      Fig. 2Traditional VMAT Plan compared to a robust VMAT Plan. (Color version of figure is available online.)
      A baseline plan was created for each patient showing the nominal dose. In robust optimization plan evaluation, parameters were defined to allow the dataset of each patient to be shifted definite numeric values isotopically. Isotropic uncertainty was set to 0.5 cm and defined as isocenter shifts placed at the endpoints of the axis in the x, y, and z directions. Isocenter shifts were also placed along the diagonals of the cuboid defined by specific uncertainties. The generated 14 scenarios were recalculated to reflect and evaluate the perturbed dose distributed representative of probabilistic setup uncertainty.

      Results

      The results for robust optimization demonstrated that the TPS was able to spread or fan the prescription dose distribution over an area along vulvar skin surface and deliver ≥ 95% of the prescription dose to the skin surface (Fig. 3). Every patient in every scenario was able to meet this metric (Table 2). Therefore, the first goal of this research was successfully achieved.
      Fig 3
      Fig. 3In robust optimization plan evaluation, parameters were defined that allowed the dataset of each patient to be shifted definite numeric values isotropically. The white dotted line is the nominal dose defined here in the DVH. The multiple magenta lines are reflective of data shifts. (Color version of figure is available online.)
      Table 2Evaluation of PTV_EVAL for planning metric V95%, 100%
      Planning Scenarios (R-L, I-S, P-A)Patient APatient BPatient CPatient DPatient E
      Baseline (0.00, 0.00, 0.00)100.00%100.00%99.99%100.00%100.00%
      Shifts (0.50, 0.00, 0.00)100.00%99.99%99.99%100.00%100.00%
      Shifts (-0.50, 0.00, 0.00)100.00%100.00%99.99%100.00%100.00%
      Shifts (0.00, 0.00, 0.50)100.00%99.99%99.98%100.00%100.00%
      Shifts (0.00, 0.00, -0.50)100.00%100.00%100.00%100.00%100.00%
      Shifts (0.00, 0.50, 0.00)100.00%100.00%99.98%100.00%100.00%
      Shifts (0.00, -0.50, 0.00)100.00%100.00%99.96%100.00%100.00%
      Shifts (0.29, 0.29, 0.29)100.00%100.00%99.99%100.00%100.00%
      Shifts (0.29, 0.29, -0.29)100.00%100.00%100.00%100.00%100.00%
      Shifts (-0.29, 0.29, 0.29)100.00%100.00%99.98%100.00%100.00%
      Shifts (-0.29, 0.29, -0.29)100.00%100.00%100.00%100.00%100.00%
      Shifts (-0.29, -0.29, -0.29)100.00%100.00%99.99%100.00%100.00%
      Shifts (0.29, -0.29, -0.29)100.00%100.00%99.99%100.00%100.00%
      Shifts (0.29, -0.29, 0.29)100.00%100.00%99.97%100.00%100.00%
      Shifts (-0.29, -0.29, 0.29)100.00%100.00%99.96%100.00%100.00%
      A, anterior; cGy, centigray; I, inferior; L, left; P, posterior; R, right; S, superior; V95%, volume receiving 95% of prescription dose.
      The results of research goal 2 where ≥ 95% of the prescription dose was delivered to 100% of the PTV_EVAL were similar to goal 1. The perturbed dose was able to fan 100% of the prescription dose over 95% the PTV_EVAL in 73 of the 75 different scenarios (Table 3). There were no obvious correlations between the direction of dataset shifts and lack of surface coverage between the 2 separate patients where scenarios failed.
      Table 3Evaluation of PTV_EVAL for planning metric V100%, 95%
      Planning Scenarios (R-L, I-S, P-A)Patient APatient BPatient CPatient DPatient E
      Baseline (0.00, 0.00, 0.00)99.90%98.23%99.21%99.96%99.31%
      Shifts (0.50, 0.00, 0.00)99.67%96.71%98.76%99.39%96.31%
      Shifts (-0.50, 0.00, 0.00)99.71%98.48%98.54%99.68%98.02%
      Shifts (0.00, 0.00, 0.50)99.24%96.70%98.29%99.56%98.64%
      Shifts (0.00, 0.00, -0.50)99.97%98.30%99.29%99.82%98.92%
      Shifts (0.00, 0.50, 0.00)99.80%96.06%
      indicates scenarios where PTV_EVAL is not meeting the metrics of V100%, 95%.
      94.73%
      99.38%99.07%
      Shifts (0.00, -0.50, 0.00)99.94%98.98%98.78%99.08%96.49%
      Shifts (0.29, 0.29, 0.29)99.24%96.05%97.38%99.33%97.84%
      Shifts (0.29, 0.29, -0.29)99.91%96.51%98.23%99.69%98.68%
      Shifts (-0.29, 0.29, 0.29)99.89%96.99%96.44%98.84%98.00%
      Shifts (-0.29, 0.29, -0.29)99.93%97.70%98.14%99.77%99.04%
      Shifts (-0.29, -0.29, -0.29)99.84%99.30%99.05%99.44%98.55%
      Shifts (0.29, -0.29, -0.29)99.99%98.35%99.18%99.22%95.81%
      Shifts (0.29, -0.29, 0.29)99.60%97.93%98.45%98.26%
      indicates scenarios where PTV_EVAL is not meeting the metrics of V100%, 95%.
      93.37%
      Shifts (-0.29, -0.29, 0.29)99.29%98.31%98.69%99.33%98.87%
      A, anterior; cGy, centi-gray, I, inferior; L, left; R, right; P, posterior; S, superior; V100%, volume receiving 100% of prescription dose.
      low asterisk indicates scenarios where PTV_EVAL is not meeting the metrics of V100%, 95%.
      Research goal 3 of delivering ≤ 110% of prescription dose to 0.03 cc of skin surface in the least desirable scenario also had positive results. Of the 75 different shifted scenarios, 70 out of 75 patients achieved the metric (Table 4). Though these patients, did not achieve the metric, the hotspots were not significantly greater with the maximum 2 doses measuring a mere 117.5%.
      Table 4Hotspot dose 0.03 cc within PTV_EVAL.
      Planning Scenarios (R-L, I-S, P-A)Patient APatient BPatient CPatient DPatient E
      5000cGy5000cGy6000cGy6765cGy5600cGy
      Baseline (0.00, 0.00, 0.00)53565319648771435849
      Shifts (0.50, 0.00, 0.00)53685342648371555827
      Shifts (-0.50, 0.00, 0.00)53435336648671485862
      Shifts (0.00, 0.00, 0.50)53785366645171955813
      Shifts (0.00, 0.00, -0.50)53595311651271845896
      Shifts (0.00, 0.50, 0.00)53645270651671415894
      Shifts (0.00, -0.50, 0.00)54335875
      indicates scenarios where hotspot of 0.03 cc exceed 110%.
      6601
      indicates scenarios where hotspot of 0.03 cc exceed 110%.
      72475859
      Shifts (0.29, 0.29, 0.29)53865314646071395835
      Shifts (0.29, 0.29, -0.29)53675301646371835881
      Shifts (-0.29, 0.29, 0.29)53565304653771285845
      Shifts (-0.29, 0.29, -0.29)53505291647971595913
      Shifts (-0.29, -0.29, -0.29)53695555
      indicates scenarios where hotspot of 0.03 cc exceed 110%.
      657171835856
      Shifts (0.29, -0.29, -0.29)53535520
      indicates scenarios where hotspot of 0.03 cc exceed 110%.
      656372105842
      Shifts (0.29, -0.29, 0.29)54055480654172515831
      Shifts (-0.29, -0.29, 0.29)54055875
      indicates scenarios where hotspot of 0.03 cc exceed 110%.
      651372145836
      A, anterior cc, cubic centimeters; cGy, centigray; I, inferior; L, left; R, right; S, superior; P, posterior.
      low asterisk indicates scenarios where hotspot of 0.03 cc exceed 110%.

      Discussion

      Patients with vulvar cancer are rarely seen in radiation oncology because of a lower incidence of vulvar cancer compared to patients presenting with more common cancers such as breast, lung, colon, and prostate. The ICRU has established safety guidelines, including TV expansions, to help create more accurate and standard methods in radiation therapy planning. During the treatment planning process, it is vital that the multileaf collimators do not close-down on the vulvar skin surface when prescription dose to the surface is necessary. This case study was performed to help demonstrate and understand the tools of robust optimization regarding surface dose to the vulvar with inadequate SMs. The newly constructed PTV_EVAL structure was the focus of the case study. The PTV_EVAL was assigned a robust factor in optimization. The datasets were shifted in a multitude of directions and data gathered was representative of setup uncertainties.
      Robust optimization demonstrated positive planning results that retrospectively looked at shifting the entire dataset in numerous directions in the controlled TPS environment. These dataset shifts hypothetically represented possible setup uncertainty. In addition to the evaluating dose coverage, hotspot dose was reviewed across the PTV_Eval structure. The average doses to the vulvar skin surface were 2.0% to 3.0% higher than traditional VMAT plans. The preventing of ≥ 110 % hotspots within the PTV_EVAL was a much harder metric to satisfy than covering the PTV_Eval. Seventy of the 75 scenarios were able to meet the desired metrics related to hotspot dose (Table 4). The correlation between not achieving this goal was dependent on the direction of the dataset shifts. The hotspot data suggested that this correlation between the hotspots increasing or not being maintained below 110% of the vulvar prescription dose was related to the patient setup uncertainty when the patient's dataset was shifted towards the gantry calculated.

      Conclusion

      Radiation therapy for treatment of vulva cancer presents several challenges. One of the most significant challenges with treating this disease is the inability to expand target volumes that adequately follow the ICRU guidelines. Traditional planning techniques, such as ISF with virtual bolus, are typically used to mitigate this problem but present new challenges for the medical dosimetrist. The purpose of this study was to determine if the VMAT treatment planning robust optimization tool could deliver a reproducible, homogenous prescription dose to the vulvar skin surface despite setup uncertainties and reduced safety margins. The goals of this research were to determine if the robust optimization technique could (1) deliver ≥ 95% of the prescription dose to 100% of the skin surface in the least desirable setup scenario, (2) deliver ≥ 100% of the prescription dose to 95% of the skin surface in the least desirable scenario and (3) not deliver ≥ 110% of the prescription dose to 0.03 cubic centimeters (cc) at the skin surface. Each of the 3 goals were attained in this case study and demonstrated the successfulness of utilizing the robust optimization planning tool for treatment of these patients.
      This case study was retrospectively completed, and future studies should include studying prospective patients in a more formal and advanced research study. Such a study would include scheduled diode readings, daily verification of patient setup, comparing patients treated with and without bolus and more detailed research that addresses the ICRU SMs. Future research should study individual variations in patient positioning, mechanical uncertainties of equipment, dosimetric uncertainties, transfer set-up errors from CT simulators to the treatment unit and human errors.

      Declaration of interests

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Acknowledgments

      The authors would like to thank Christopher Ward, CMD, for his contributions and knowledge with Raystation TPS. The primary author would like to thank the medical staff at Swedish Cancer Institute, Seattle, WA for advocating higher education and the use of the computer software used to create this case study.

      References

        • Weinberg D
        • Gomez-Martinez RA.
        Vulvar cancer.
        Obstet Gynecol Clin N Am. 2019; 46: 125-135https://doi.org/10.1016/j.ogc.2018.09.008
        • Library of Congress Cataloging-in-Publication Data
        International Commission on Radiation Units and Measurements. ICRU Report 62.
        Prescribing, Recording and Reporting Photon Beam Therapy (supplement to ICRU 50), Bethesda, Maryland1999
        • Dyer BA
        • Jenshus A
        • Mayadev JS.
        Integrated skin flash planning technique for intensity-modulated radiation therapy for vulvar cancer prevents marginal misses and improves superficial dose coverage.
        Med Dosim. 2019; 44: 7-10