If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Skin tattoos represent the standard approach for surface alignment and setup of breast cancer radiation therapy, yet permanent skin markings contribute to adverse cosmesis and patient dissatisfaction. With the advent of contemporary surface-imaging technology, we evaluated setup accuracy and timing between “tattoo-less” and traditional tattoo-based setup techniques.
Methods and Materials
Patients receiving accelerated partial breast irradiation (APBI) underwent traditional tattoo-based setup (TTB), alternating daily with a tattoo-less setup via surface imaging using AlignRT (ART). Following initial setup, position was verified via daily kV imaging, with matching on surgical clips representing ground truth. Translational shifts (TS) and rotational shifts (RS) were ascertained, as were setup time and total in-room time. Statistical analyses used the Wilcoxon signed rank test and Pitman-Morgan variance test.
Results
A total of 43 patients receiving APBI and 356 treatment fractions were analyzed (174 TTB fractions and 182 using ART). For tattoo-less setup via ART, the median absolute TS were 0.31 cm in the vertical (range, 0.08-0.82), 0.23 cm in the lateral (0.05-0.86), and 0.26 cm in the longitudinal (0.02-0.72) axes. For TTB setup, the corresponding median TS were 0.34 cm (0.05-1.98), 0.31 cm (0.09-1.84), and 0.34 cm (0.08-1.25), respectively. The median magnitude shifts were 0.59 (0.30-1.31) for ART and 0.80 (0.27-2.13) for TTB. ART was not statistically distinguishable from TTB in terms of TS, except in the longitudinal direction (P = .154, .059, and .021, respectively), and was superior to TTB for magnitude shift (P < .001). The variance of each TS variable was significantly narrower for ART compared with TTB (P ≤ .001 vertical, P = .001 lateral, P = .005 longitudinal). The median absolute RS for ART was 0.64° rotation (range, 0.00-1.90), 0.65° roll (0.05-2.90), and 0.30° pitch (0.00-1.50). The corresponding median RS for TTB were 0.80° (0.00-2.50), 0.64° (0.00-3.00), and 0.46° (0.00-2.90), respectively. ART setup was not statistically different from TTB in terms of RS (P = .868, .236, and .079, respectively). ART showed lower variance than TTB in terms of pitch (P = .009). The median total in-room time was shorter for ART than TTB (15.42 vs 17.25 minutes; P = .008), as was the median setup time (11.12 vs 13.00 minutes; P = .001). Moreover, ART had a narrower distribution of setup time with fewer lengthy outliers versus TTB.
Conclusions
These findings suggest that a tattoo-less setup approach with AlignRT may be sufficiently accurate and expeditious to supplant surface tattoos for patients receiving APBI. Further analyses with larger cohorts will determine whether tattoo-based approaches can be replaced by noninvasive surface imaging.
Introduction
Skin tattoos have long been a standard component of contemporary radiation therapy (RT) setup procedures.
These permanent skin markings represent visible index points on patients receiving RT that are aligned with precisely calibrated in-room lasers to position patients for treatment. This alignment facilitates positioning of the patient at precise coordinates in the treatment room, permitting accurate and reproducible treatment delivery. While various techniques have been described, setup for breast RT may require up to 12 tattoos depending on the clinical scenario.
In recent years, accelerated partial breast irradiation (APBI) has shown safety and feasibility for appropriately selected patients with breast cancer. Patient setup for APBI is typically initiated with conventional tattoo-based (TTB) alignment procedures, subsequently followed by megavolt or kV x-ray image guidance at each fraction to confirm patient positioning and target location. Image guidance is subsequently based on visualization of the postoperative seroma or on surgical clips that define the postlumpectomy tumor bed and target volume.
While TTB setup has been standard since the advent of contemporary RT, a large and growing body of literature suggests that patients dislike this approach. Indeed, patients report negative feelings toward tattoos based on cultural and religious beliefs, discomfort from the tattooing procedure, effect on body image and self-esteem, and the psychological distress of bearing a permanent reminder of their cancer diagnosis.
With the development of surface-imaging technologies, such as AlignRT, skin tattoos may be obviated by the ability to visualize and setup a relevant patient surface in its entirety. Here, we sought to compare the setup performance of tattoos versus AlignRT for patients with breast cancer receiving APBI. Using image guidance of the tumor bed/surgical clips as ground truth for target localization, shifts in 6° of freedom were compared between TTB and AlignRT setups, as were in-room and setup times. We hypothesized that use of AlignRT would be at least as accurate and rapid as TTB setup.
Methods and Materials
Consecutive patients with early-stage breast cancer who were to undergo APBI at our center were identified. Simulation and treatment planning were conducted as previously described.
Briefly, all patients underwent breast conservation surgery, during which surgical clips were placed in the resection cavity for delineation of the tumor bed. Computed tomography (CT) simulation was conducted in the supine position using a breast board with both arms abducted. Four tattoos were placed per institutional practice: (1) at the level of the nipple in the midline, (2) at the level of the nipple in the midaxillary line, (3) 10 cm inferior to the level of the nipple in the midline, and (4) 10 cm inferior to the level of the nipple in midaxillary line (Fig. 1). Using the CT simulation images, the radiation oncologist outlined the tumor bed by noting contrasting areas of density as well as the presence of clips in the breast. A margin of 1.5 to 2 cm was added isotropically to create the planning target volume (PTV), which was limited anteriorly to 5 mm below the skin surface and posteriorly to the anterior surface of the ribs. Surgical clips were contoured to facilitate daily image guidance. Typical treatment plans comprised 3-dimensional (3D) conformal RT using 3 beams (2 “mini”-tangent beams and an en face electron beam) optimized to administer homogeneous dose throughout the PTV and to mitigate hotspots.
Figure 1(A) Workflow of alternating tattoo-based and AlignRT setup procedures on alternating days. (B) Sample skin-marking tattoo.
Following treatment planning, patients underwent once daily APBI to the surgical bed to a total dose of 40 Gy in 10 fractions (ie, 4 Gy per fraction). To compare TTB versus surface-imaging setup approaches for each patient, techniques were alternated daily for every patient: traditional TTB was employed for all APBI patients on odd-numbered days of the month, alternating with tattoo-less setup via surface imaging using AlignRT (ART) on even days of the month (Fig. 1). Treating radiation therapists recorded the times at which a patient entered the treatment room, when setup began, when treatment was initiated, and when the patient exited the treatment room. Fractions with any missing data (ie, in-room time, setup time, translational shifts [TS], or rotational shifts [RS]) were excluded from the analysis.
TTB setup comprised (1) initial positioning with laser alignment to the tattoos and (2) treatment target verification via orthogonal kV x-rays with 2-dimensional matching on surgical clips. ART setup comprised (1) initial positioning with 3D surface imaging via AlignRT (lasers were off, and tattoos were not considered during this procedure) and (2) treatment target verification via orthogonal kV x-rays matching on surgical clips. AlignRT version 6.2 was used with a 3-pod system along with postural alignment and 3-video module for postural corrections.
Following the initial daily setup for either TTB or ART, ground truth was based on position verification with daily kV imaging matching on surgical clips (via the Varian [Palo Alto, CA] TrueBeam OBI kV imaging system using the “Thorax Arms Up” imaging protocol). Lumpectomy bed clips were projected onto the reference digitally reconstructed radiograph (DRR). Image registration was performed by aligning surgical bed clips from the daily kV x-rays to the reference DRR clips. Automatic kV image registration was used and verified manually by the treatment radiation therapist. Manual corrections were employed if automated alignment was unsatisfactory. Any additional setup corrections from x-ray imaging after either setup approach were recorded. Couch shifts in 6° of freedom were employed to match the image guided position of surgical clips to their position in the treatment plan DRR. TS and RS were ascertained, as were setup time and total in-room time.
Surface imaging for ART employed a reference image generated from the CT simulation scan with the patient in the treatment position. The region of interest was defined as the involved breast (including the PTV with adequate surrounding tissue).
Statistical methods
The absolute values for TS and RS for each fraction were recorded, and the mean of these values for a given patient was compared as stratified by the setup approach (ie, the mean of all TTB TS for each patient were compared with the mean of all ART TS). Because the resultant means were not normally distributed, we applied nonparametric analysis methods. To evaluate accuracy, we employed the Wilcoxon signed rank test on paired, nonparametric data, testing differences in the median of the means between groups (hypothesizing that the median difference). To evaluate precision, we employed the Pitman-Morgan test to assess differences in the variances between means in each setup group.
Data representing extreme outliers for any measurements, defined as beyond the bounds of the first quartile minus 1.5 interquartile ranges (IQR, Q1 – 1.5) or the third quartile plus 1.5 (IQR, Q3 + 1.5), were classified as measurement or data entry errors and were omitted from the analysis. The threshold for significance on all statistical tests was set at P < .05. Analyses were conducted using R, version 4.2.2 (R Project for Statistical Programming, Vienna, Austria).
Results
APBI was administered to 43 consecutive patients using alternating daily TTB and ART setup procedures as described in Methods and Materials, yielding a total of 356 analyzable fractions (174 TTB, 182 ART). Seventy fractions were excluded due to incomplete data collection or unevaluable data entry.
Along the vertical (VRT), lateral (LAT), and longitudinal (LNG) axes, respectively, mean shifts were 0.44 cm for TTB versus 0.34 cm for ART (VRT; P = .154), 0.39 cm TTB versus 0.29 cm ART (LAT; P = .059), and 0.40 cm TTB versus 0.28 cm ART (LNG; P = .021) (Table 1, Fig. 2A). Variance testing of each TS axis showed a narrower distribution of observations with ART versus TTB (P ≤ .001, P ≤ .001, and P ≤ .005, respectively) (Table 2).
Table 1Comparison of translational shifts and timing between tattoo-based and AlignRT-based setup techniques
Tattoo (N = 43, f = 174)
AlignRT (N = 43, f = 182)
Difference (N = 43)
P value
Average in-room time (min)
.005
Mean (SD)
17.32 (4.02)
15.62 (3.32)
–1.69 (3.45)
Median (Ql, Q3)
17.25 (14.73, 19.67)
15.42 (13.17, 17.50)
–1.08 (–3.92, 0.58)
Min-max
9.50-26.00
9.80-22.55
–8.67 to 5.50
Average setup time (min)
<.001
Mean (SD)
13.46 (4.02)
11.50 (3.58)
–1.96 (3.55)
Median (Ql, Q3)
13.00 (10.66, 15.56)
11.12 (8.98, 13.89)
–2.01 (–4.50, –0.17)
Min-max
5.66-21.60
5.58-19.95
–9.42 to 5.61
Absolute vertical axis shift (cm)
.154
Mean (SD)
0.44 (0.35)
0.34 (0.19)
–0.10 (0.33)
Median (Ql, Q3)
0.34 (0.23, 0.53)
0.31 (0.20, 0.44)
–0.02 (–0.21, 0.08)
Min-max
0.05-1.98
0.08-0.82
–1.35 to 0.36
Absolute lateral axis shift (cm)
.059
Mean (SD)
0.39 (0.31)
0.29 (0.18)
–0.10 (0.36)
Median (Ql, Q3)
0.31 (0.23, 0.48)
0.23 (0.16, 0.42)
–0.06 (–0.21, 0.08)
Min-max
0.09-1.84
0.05-0.86
–1.71 to 0.59
Absolute longitudinal axis shift (cm)
.021
Mean (SD)
0.40 (0.25)
0.28 (0.16)
–0.13 (0.32)
Median (Ql, Q3)
0.34 (0.22, 0.52)
0.26 (0.16, 0.36)
–0.09 (–0.28, 0.03)
Min-max
0.08-1.25
0.02-0.72
–1.16 to 0.38
Magnitude shift (cm)
<.001
Mean (SD)
0.86 (0.39)
0.62 (0.23)
–0.23 (0.41)
Median (Ql, Q3)
0.80 (0.59, 1.01)
0.59 (0.45, 0.75)
–0.20 (–0.34, –0.02)
Min-max
0.27-2.13
0.30-1.31
–1.49 to 0.40
Abbreviations: max = maximum; min = minimum; Q = quarter; SD = standard deviation.
Wilcoxon signed rank test.
A significant result (P < .05) indicates AlignRT and tattoo-based setup techniques perform differently within that domain.
Statistics to compare: median of the means. A negative result in difference is preferrable to AlignRT.
Magnitude is defined as the geometric average of vertical, lateral, and longitudinal shifts.
With regard to the rotational axes of yaw (RTN), roll, and pitch, respectively, mean shifts were 0.89° for TTB versus 0.80° for ART (RTN; P = .868), 0.82° for TTB versus 0.74° for ART (roll; P = .236), and 0.64° for TTB versus 0.44° for ART (pitch; P = .079) (Table 3, Fig. 2B). Variance testing of rotational shift variables are shown in Table 4. ART showed lower variance than TTB in terms of pitch (P = .009; Table 4).
Table 3Comparison of rotational shifts between tattoo-based and AlignRT-based setup techniques
Tattoo (n = 31, f = 108)
AlignRT (n = 31, f = 114)
Difference (n = 31)
P value
Absolute rotational axis shift (degree)
.868
Mean (SD)
0.89 (0.69)
0.80 (0.54)
–0.09 (0.65)
Median (Ql, Q3)
0.80 (0.38, 1.23)
0.64 (0.35, 1.16)
0.02 (–0.30, 0.28)
Min-max
0.00-2.50
0.00-1.90
–2.30 to 0.88
Absolute roll shift (degree)
.236
Mean (SD)
0.82 (0.71)
0.74 (0.67)
–0.08 (0.83)
Median (Ql, Q3)
0.64 (0.40, 1.07)
0.65 (0.21, 1.09)
–0.07 (–0.59, 0.22)
Min-max
0.00-3.00
0.05-2.90
–1.40 to 2.90
Absolute pitch shift (degree)
.079
Mean (SD)
0.64 (0.70)
0.44 (0.43)
–0.20 (0.52)
Median (Ql, Q3)
0.46 (0.18, 0.87)
0.30 (0.17, 0.51)
–0.10 (–0.46, 0.08)
Min-max
0.00-2.90
0.00-1.50
–1.47 to 0.70
Abbreviations: max = maximum; min = minimum; Q = quartile; SD = standard deviation.
Wilcoxon signed rank test.
A significant result (P < .05) indicates AlignRT and tattoo-based perform differently within that domain.
Statistics to compare: median of the means. A negative result in difference is preferrable to AlignRT.
Rotational shifts are only recorded in latest AlignRT version (version 6.2).
Considering that the 2 setup techniques might exhibit workflow or setup efficiency differences, we evaluated total in-room time and setup time on each treatment day (Table 1, Fig. 3). The median total in-room time for the administered APBI treatments with TTB was 17.25 minutes (IQR, 14.73-19.67) versus 15.42 minutes with ART (IQR, 13.17-17.50; P = .005 favoring ART). To determine whether this difference was attributable to the actual efficiency of patient setup or to facility with utilization of the equipment, setup time was also evaluated. Median setup time for TTB was 13 minutes (IQR, 10.66-15.56) versus 11.12 minutes for ART (IQR, 8.98-13.89; P < .001). Moreover, ART exhibited a narrower distribution of setup time with fewer lengthy extremes compared with TTB.
Figure 3Average (A) in-room and (B) setup time distribution.
These findings, comprising the largest number of comparative measurements in the largest cohort to date, demonstrate that among patients receiving APBI, surface imaging with daily image guidance is similarly accurate and significantly faster than TTB setup. Evaluation of 3 TS and 3 RTN axes revealed no dimension in which tattoos were more accurate than surface-guided imaging (ie, AlignRT), while time measurements demonstrated significantly faster setup and less in-room time with surface-guided imaging.
Fractionated RT has depended upon patient positioning via permanent surface markings for over half a century.
Among patients with breast cancer, perennial concerns regarding cosmesis and emotional distress routinely call into question the necessity of TTB setup practices. Several alternatives to conventional tattoos have emerged in recent decades, including fluorescent (invisible) ink tattoos, surface markers, and water-based soluble tattoos.
These methods all serve to obviate the visibly permanent marking of conventional tattoos, yet retain the limitations of setting up to a limited number of spatial points adjacent to an otherwise deformable structure (ie, the breast) in which anatomy may vary depending on daily positioning and body habitus. In a recent survey of relevant patients, 70% reported negative feelings about their setup tattoos, and 78% would have chosen a more cumbersome treatment if tattoos could be avoided.
Notably, to avoid tattoos, survey respondents were willing to travel an additional 45.5 miles, over an additional 39.3 minutes, at an added expense of $37.50 per trip.
To this end, surface-guided RT has garnered increasing attention in recent decades. A pioneering study of AlignRT among 20 patients receiving whole breast RT concluded that surface imaging was helpful above and beyond tattoos, but that an additional imaging modality was also needed to establish ground truth (in this case, daily cone beam CT).
This study followed a similar analysis of 50 patients receiving whole breast RT with 1258 analyzable treatment fractions showing that daily alignment with 3D surface imaging reduced setup errors compared with skin marks alone, and that skin marks were particularly poor at setting up the antero-posterior axis.
In another whole breast RT study, performed in the United Kingdom, 43 patients undergoing radiation to the breast or chest wall were evaluated to compare traditional TTB setup with AlignRT surface-guided setup.
Mean shifts in this study showed that tattoo-less setup with SGRT was comparable to TTB setup for right-sided breast treatments and superior for left-sided breast treatments using deep inspiration breath hold.
The more recent advent of partial breast radiation brought increased urgency to improvements in patient setup procedures. In light of the narrower margins used for APBI in comparison with whole breast RT, investigators proposed that tighter APBI tolerances to reduce systematic error could narrow PTV margins and improve morbidity outcomes. Among the first of these studies was evaluation of an early stereo photogrammetry system in 9 patients over 53 fractions at the Massachusetts General Hospital, showing preliminarily that surface imaging was superior to setup using skin marks and lasers.
Approximately a decade later, the same group conducted a pilot study of tattoo-less partial breast irradiation, evaluating 10 patients who underwent positioning via surface imaging without reference to tattoos.
This analysis was largely consistent with our own observations: setup time per fraction was approximately 2 minutes shorter with surface imaging, and no significant differences were observed in 3 TS. This pilot study concluded that surface imaging with matching to surgical clips could obviate the need for skin-based tattoos in patients receiving APBI. Building on these findings, further studies are underway to evaluate the possibility of omitting tattoos even among those receiving regional nodal irradiation, with consistent advantages for surface-guided imaging in early reports.
Though several small studies have undertaken to compare surface-guided imaging with traditional TTB setup, our study represents the largest cohort to date and is the only to report setup shifts in 3 RTN axes in addition to the 3 TS axes. Moreover, other studies have universally compared 2 independent cohorts (ie, tattoo vs tattoo-less), whereas our approach used each patient as their own internal control by alternating tattoo versus tattoo-less days, thereby mitigating potential sources of interpatient systematic bias which may be particularly problematic among smaller cohorts.
Our results must be interpreted in the context of the analysis design. Importantly, we collected no data on patient reported outcomes or perceptions of the 2 setup approaches. While existing literature and anecdotal evidence suggest that, on balance, patients feel negatively toward tattoos, the magnitude of this effect could not be analyzed in our cohort. Indeed, because all patients in our cohort received tattoos for setup on TTB days, we could not ascertain patient satisfaction in the absence of tattoos. Along similar lines, while all patients had skin tattoos that were available to the therapy teams on ART days as a potential source of confounding, in-room lasers were disabled during setup on these days such that visible surface markings could not have influenced patient alignment. It is also important to note that this analysis only reflects our observations regarding setup for APBI. Extrapolation to whole-breast irradiation or regional nodal irradiation will require further study, considering the larger treatment volumes targeted by these approaches, the frequent absence of tumor-bed clips on which to match kV images in the postmastectomy setting, and the use of a variety of motion management and breath hold techniques in these settings, which may alter the accuracy or reproducibility of surface imaging.
Conclusion
Our findings suggest that a setup procedure based on surface-guided imaging and daily matching on clips may be sufficiently accurate and efficient to supplant skin tattoos for patients receiving APBI. Outcomes analyses among those not receiving tattoos will elucidate the efficacy and desirability of this less invasive treatment approach at improving setup accuracy, efficiency, and patient quality of life.
Acknowledgments
The authors acknowledge Andrea Nevling, Ariana Yager, Chrissy Gross, Courtney Vasta, Danielle Scrazati, Deepali Patel, Erin Terrafranca, James Rodgers, Jason Cordero, Kimberly Maura, Laura Nappi, Lauren Breen, Maria Lizarzaburu, Raizy Streicher, Shelley Davis, Steven Leung, and Taylor Kornegay for operational assistance. They thank Anthony Abaya and Michael Sullivan for support with data management and research operations.
References
Greer PB
Mortensen TM.
Anterior-posterior treatment localization in pelvic radiotherapy: Tattoos or fixed couch-to-isocentre distance.
Presented at the 2022 European Society of Radiation Oncology Annual Meeting, May 6-10, 2022.
Sources of support: The preparation of this study was supported in part by the Lois Green Fund, the Rose-Margulies Family Research Fund, and National Institutes of Health/National Cancer Institute Cancer Center Support Grant No. P30CA008748.
Disclosures: 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.
Interested parties may contact the corresponding author to request access to anonymized study data.
Pitman-Morgan variance test.Statistics in use: variance ratio. A ratio >1 indicates preference to AlignRT, while <1 indicates the opposite.
00012-X&id=gr1.jpg){kind=link}
00012-X&id=gr2.jpg){kind=link}
00012-X&id=gr3.jpg){kind=link}