Abstract
Despite advances in treatment and response assessment in locally advanced rectal cancer (LARC), it is unclear which patients should undergo nonoperative management (NOM). We performed a single-center, retrospective study to evaluate post-total neoadjuvant therapy (TNT) circulating tumor DNA (ctDNA) in predicting treatment response. We found that post-TNT ctDNA had a sensitivity of 23% and specificity of 100% for predicting residual disease upon resection, with a positive predictive value (PPV) of 100% and a negative predictive value (NPV) of 47%. For predicting poor tumor regression on MRI, ctDNA had a sensitivity of 16% and specificity of 96%, with a PPV of 75% and NPV of 60%. A commercially available ctDNA assay was insufficient to predict residual disease after TNT and should not be used alone to select patients for NOM in LARC.
Keywords: circulating tumor DNA, rectal cancer, total neoadjuvant therapy, nonoperative management
Despite advances in treatment and response assessment in locally advanced rectal cancer, it is unclear which patients should undergo nonoperative management. This article evaluates post-total neoadjuvant therapy circulating tumor DNA in predicting treatment response.
Introduction
Treatment for locally advanced rectal cancer (LARC) has changed significantly in the past decade, with all (chemo)radiation and chemotherapy moved prior to surgery in total neoadjuvant therapy (TNT), now standard of care.1 With improved disease control up-front, there is interest in sparing surgical morbidity using nonoperative management (NOM).1 Similar to how circulating tumor DNA (ctDNA) is used in stage II/III colorectal cancer to spare adjuvant chemotherapy, this biomarker is being investigated in LARC to assess candidacy for NOM.2-6 Herein, we investigate ctDNA as a predictor of pathologic, radiographic, and endoscopic response after TNT to determine its utility in selecting patients for NOM.
Methods
In this single-center, retrospective study at the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University, we included all patients in our rectal cancer clinic 18 or older with LARC (cT3-T4NxM0 or cTxN1-2M0), confirmed by imaging and histology, representative of our center’s general population. Data were collected from February 2021 to May 2023. All patients received TNT with short-course radiotherapy (RT) with 25 Gy in 5 sessions (5 Gy/session) plus 6 cycles of capecitabine and oxaliplatin (CAPOX) or 9 cycles of leucovorin calcium, fluorouracil, and oxaliplatin (FOLFOX), or long-course RT with capecitabine plus 4-6 cycles of CAPOX or 6-9 cycles of FOLFOX, except one patient planned to receive 9 cycles of FOLFOX who received 3 due to issues with follow-up. We excluded patients undergoing treatment for locally recurrent or metastatic disease, microsatellite instability-high disease, and without post-TNT ctDNA.
Magnetic resonance imaging (MRI) was performed in all patients (median: 1.1, range: −5.4 to 13.3 weeks from completion), with MRI tumor regression grade (mrTRG) determined by radiologists at our institution. A complete/good radiographic response was mrTRG1-2, with poor response mrTRG3-5. ctDNA was collected at multiple timepoints: pretreatment (median: 9, range: −31 to 0 days), on-treatment (median: 7.3, range: 2.4-11.9 weeks after beginning treatment), post-treatment (median: 1.1, range: −6.6 to 32.1 weeks from completion), and postoperatively (within 8 weeks following surgery). Post-TNT endoscopic assessment was performed in 41 patients (93%; median: 1.4, range: −1.0 to 14.9 weeks from completion), with complete clinical response (cCR) or residual disease determined by performing physician(s). All ctDNA levels were determined using Natera’s Signatera Residual Disease Test, a tumor-informed ctDNA assay incorporating mutational profiles from each tumor to target patient-specific variants.
Continuous variables were described with median, range, number, and percentage, with Wilcoxon test to compare median values. Differences among categorical variables were assessed using Fisher’s exact test. P-value of < .05 was statistically significant.
Results
Our cohort included 44 patients: 22 (50%) undergoing surgery, 19 (45%) opting for NOM, and 3 (5%) awaiting surgery at cutoff. When comparing operative patients with those undergoing NOM, we found a significant difference in cT stage, mrTRG, and endoscopic response after TNT, with a trend toward a difference in distance from the anal verge (Table 1). The median time between RT and surgery was 200 days (range: 50-413 days). Of 22 patients undergoing surgery, 9 (41%) had a pathologic complete response (pCR), of whom 2 had complete/good response on MRI and one had cCR on endoscopy (Table 1). Full demographic and response data are in Table 1.
Table 1.
Patient demographics and response, comparing surgery with non-operative management.a
| Characteristic | Overall (n = 44) | Surgeryb (n = 25) | Non-operative management (n = 19) | P-value |
|---|---|---|---|---|
| Age on study, median (range) | 56 (32-80) | 57 (32-80) | 56 (32-75) | 1.00 |
| Sex, female, no. (%) | 18 (41) | 11 (44) | 7 (39) | .76 |
| Performance status, no. (%) | ||||
| 0 | 35 (80) | 18 (72) | 17 (89) | .26 |
| 1 | 9 (20) | 7 (28) | 2 (11) | |
| Race, no. (%) | ||||
| Caucasian | 34 (77) | 19 (76) | 15 (79%) | 1.00 |
| African American | 4 (9) | 2 (8) | 2 (11%) | |
| Other | 6 (14) | 4 (16) | 2 (11%) | |
| Regimen, no. (%) | ||||
| SCRT + chemotherapy | 40 (91) | 22 (88) | 18 (95) | .62 |
| LCRT + chemotherapy | 4 (9) | 3 (12) | 1 (5) | |
| Distance from anal verge, no. (%) | ||||
| High (10.1-15 cm) | 10 (23) | 9 (36) | 1 (5) | .05 |
| Mid (5.1-10 cm) | 19 (43) | 9 (36) | 10 (53) | |
| Lower (0-5 cm) | 15 (34) | 7 (28) | 8 (42) | |
| cT stage, no. (%) | ||||
| 1-2 | 7 (16) | 1 (4) | 6 (32) | .03 |
| 3-4 | 37 (84) | 24 (96) | 13 (68) | |
| cN stage, no. (%) | ||||
| 0-1 | 29 (66) | 17 (68) | 12 (63) | .75 |
| >2 | 15 (34) | 8 (32) | 7 (37) | |
| mrTRG response, no. (%) | ||||
| TRG1-2 | 25 (57) | 7 (28) | 18 (95) | <.01 |
| TRG3-5 | 19 (43) | 18 (72) | 1 (5) | |
| Endoscopic response | ||||
| Overall, no. | 41 | 23 | 18 | |
| cCR, no. (%) | 22 (54) | 6 (26) | 16 (89) | <.01 |
| Residual disease, no. (%) | 19 (46) | 17 (74) | 2 (11)c | |
| Pathologic response, no. (%) | ||||
| pCR | N/A | 9 (41) | N/A | N/A |
| Residual disease | N/A | 13 (59) | N/A | |
aSelect percentages may not add up to 100 due to rounding.
bThree patients included in the surgery category were awaiting surgery at the time of data cutoff so did not have pathologic response data available.
cBoth patients with residual disease had subsequent endoscopic assessment within the following 6 months that demonstrated complete clinical response, although one patient did have a tubular adenoma lateral to the location of his/her rectal cancer on subsequent assessments.
Abbreviations: SCRT, short-course radiotherapy; LCRT, long-course radiotherapy; cT, clinical tumor stage; cN: clinical lymph node stage; mrTRG, magnetic resonance tumor regression grade; cCR, complete clinical response; pCR, pathologic complete response.
Among 22 patients with pre-treatment ctDNA, levels were detectable in 18 (82%) prior to TNT. While ctDNA levels decreased in all patients (representative ctDNA trends shown in Fig. 1), ctDNA remained detectable in 3 (13%) after TNT (Table 2). Post-TNT ctDNA had a sensitivity of 23% and specificity of 100% for predicting residual disease upon resection, with a positive predictive value (PPV) of 100% and a negative predictive value (NPV) of 47% (Table 2). In evaluating ctDNA to predict mrTRG, post-TNT ctDNA had a sensitivity of 16% and specificity of 96% in predicting poor tumor regression, with PPV 75% and NPV 60%. All patients with complete/good response on imaging had undetectable ctDNA after TNT. MRI had a sensitivity of 62% and specificity of 22%, with PPV 53% and NPV 29% for predicting residual disease on resection after TNT. In evaluating ctDNA to predict endoscopic response, post-TNT ctDNA had a sensitivity of 5% and specificity of 95% in predicting residual disease, with PPV 50% and NPV 54%. One patient with cCR on endoscopy had detectable ctDNA post-TNT. Endoscopy had a sensitivity of 58% and specificity of 11%, with PPV 47% and NPV 17% to predict residual disease upon resection.
Figure 1.
ctDNA trends throughout treatment. Eleven patients had ctDNA values available at multiple time points (the pre-treatment, on-treatment, post-treatment, and/or post-operative time points). Patients with only undetected ctDNA values on-treatment, post-TNT, and/or post-surgery were excluded. Patients with a pCR shown in yellow and patients without a pCR shown in blue. The y-axis, or timepoint zero, represents the start of TNT, while the red triangle represents that end of TNT for each patient.
Table 2.
Patient demographics and response based on ctDNA pre- and post-TNT.a
| Characteristic | Pre-TNT ctDNA (n = 22) | Post-TNT ctDNA (n = 44) | ||||
|---|---|---|---|---|---|---|
| Positive (n = 18) | Negative (n = 4) | P-value | Positive (n = 4) | Negative (n = 40) | P-value | |
| Age on study, median (range) | 57 (43-76) | 59 (49-72) | 1.00 | 61 (57-80) | 55 (32-76) | 1.00 |
| Sex, female, no. (%) | 6 (33) | 2 (50) | .60 | 2 (50) | 16 (40) | 1.00 |
| Performance status, no. (%) | ||||||
| 0 | 14 (78) | 3 (75) | 1.00 | 1 (25) | 34 (85) | .02 |
| 1 | 4 (22) | 1 (25) | 3 (75) | 6 (15) | ||
| Race, no. (%) | ||||||
| Caucasian | 13 (72) | 4 (100) | 1.00 | 4 (100) | 30 (75) | 1.00 |
| African American | 2 (11) | 0 (0) | 0 (0) | 4 (10) | ||
| Other | 3 (17) | 0 (0) | 0 (0) | 6 (15) | ||
| Regimen, no. (%) | ||||||
| SCRT + chemotherapy | 17 (94) | 4 (100) | 1.00 | 1 (25) | 37 (8) | .33 |
| LCRT + chemotherapy | 1 (6) | 0 (0) | 3 (75) | 3 (92) | ||
| Distance from anal verge, no. (%) | ||||||
| High (10.1-15 cm) | 4 (22) | 0 (0) | .80 | 1 (25) | 9 (23) | .81 |
| Mid (5.1-10 cm) | 6 (33) | 2 (50) | 2 (50) | 18 (45) | ||
| Lower (0-5 cm) | 8 (44) | 2 (50) | 1 (25) | 13 (33) | ||
| cT stage, no. (%) | ||||||
| 1-2 | 1 (6) | 1 (25) | .34 | 0 (0) | 7 (18) | 1.00 |
| 3-4 | 17 (94) | 3 (75) | 4 (100) | 33 (82) | ||
| cN stage, no. (%) | ||||||
| 0-1 | 13 (72) | 2 (50) | .56 | 3 (75) | 26 (65) | 1.00 |
| >2 | 5 (28) | 2 (50) | 1 (25) | 14 (35) | ||
| mrTRG response, no. (%) | ||||||
| TRG1-2 | 9 (50) | 4 (100) | .12 | 1 (25) | 24 (60) | .30 |
| TRG3-5 | 9 (50) | 0 (0) | 3 (75) | 16 (40) | ||
| Endoscopic response | ||||||
| Overall, no. | 16 | 4 | 2 | 39 | ||
| cCR, no. (%) | 8 (50) | 3 (75) | .59 | 1 (50) | 21 (54) | 1.00 |
| Residual disease, no. (%) | 8 (50) | 1 (25) | 1 (50) | 18 (46) | ||
| Pathologic response | ||||||
| Overall, no. | 9 | 0 | 3 | 19 | ||
| pCR, no. (%) | 2 (22) | 0 (N/A) | N/A | 0 (0) | 9 (47) | .24 |
| Residual disease, no. (%) | 7 (78) | 0 (N/A) | 3 (100) | 10 (53) | ||
aSelect percentages may not add up to 100 due to rounding.
Abbreviations: TNT, total neoadjuvant therapy; SCRT, short-course radiotherapy; LCRT, long-course radiotherapy; cT, clinical tumor stage; cN, clinical lymph node stage; mrTRG, magnetic resonance tumor regression grade.
Discussion
Given the complications associated with operative management in LARC, including fecal incontinence, increased stool frequency, urgency, and sexual dysfunction, as well as improved disease control with TNT, NOM has become an attractive option. Unfortunately, conventional tools, including MRI, endoscopy, and physical examination, often overestimate treatment response and underestimate residual disease, making molecular testing with ctDNA an attractive option to enhance detection.1,7 In this real-world population with LARC, a commercially available ctDNA assay had poor sensitivity (23%) and NPV (47%) in detecting residual disease after TNT, suggesting that ctDNA alone is insufficient to determine who should undergo NOM.3,4 These findings are consistent with previous studies evaluating ctDNA to predict pCR after neoadjuvant chemoradiotherapy, where a significant portion of patients with negative post-treatment ctDNA had residual disease on resection.3,5,8 To our knowledge, the first prospective studies to ask this question are GEMCAD-REVEAL, which will evaluate NPV and PPV of post-TNT ctDNA to predict relapse in LARC, as well as the JANUS Rectal Cancer Trial, which will correlate ctDNA with radiographic, pathologic, and clinical outcomes in an intensified TNT regimen.9,10 Until their completion, however, we have to rely on retrospective studies to guide ctDNA use in this setting.
Limitations of this study include small sample size and limited long-term follow-up, which precluded survival analysis. Given the large number of patients in the NOM arm, the sensitivity and NPV of ctDNA for detecting residual disease may be higher than that shown, as the surgery group had a higher proportion of patients with poor response on MRI, residual disease on endoscopy, and T3-T4 tumors and, thus, were less likely to have pCR. Studies in larger cohorts with higher sensitivity assays are needed before ctDNA can be used to select patients for NOM in LARC.
Contributor Information
Stephanie L Alden, Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, USA.
Valerie Lee, Department of Oncology, Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Johns Hopkins University School of Medicine, Baltimore, MD, USA.
Amol K Narang, Department of Radiation Oncology & Molecular Radiation Sciences, Sidney Kimmel Cancer Center at Johns Hopkins, Johns Hopkins University School of Medicine, Baltimore, MD, USA.
Jeffrey Meyer, Department of Radiation Oncology & Molecular Radiation Sciences, Sidney Kimmel Cancer Center at Johns Hopkins, Johns Hopkins University School of Medicine, Baltimore, MD, USA.
Susan L Gearhart, Department of Surgery, Johns Hopkins Bayview Medical Center, Baltimore, MD, USA.
Eric S Christenson, Department of Oncology, Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Johns Hopkins University School of Medicine, Baltimore, MD, USA.
Funding
All authors report support from the Bloomberg-Kimmel Institute for Cancer Immunotherapy and Cancer Convergence Institute. Dr. Amol K. Narang reports research funding from Boston Scientific, Flavocure, and Nanocan. Dr. Jeffrey Meyer reports previous funding from Boston Scientific, which ended in December 2021. Dr. Eric Christenson reports research collaborations with Haystack, Pfizer, Affirmed, Regeneron, and NextCure, as well as support from Swim Across America. The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Conflict of Interest
Amol K. Narang has received research funding from Boston Scientific, Flavocure, and Nanocan. Jeffrey Meyer has received royalties from UpToDate and Springer. Eric S. Christenson has performed consulting for Seres Therapeutics and SIRTeX and has research collaborations with Haystack, Pfizer, Affirmed, Regeneron, and NextCure. The other authors indicated no financial relationships.
Data Availability
The data used to create this article will be shared on reasonable request to the corresponding author.
References
- 1. Zwart WH, Hotca A, Hospers GAP, Goodman KA, Garcia-Aguilar J.. The multimodal management of locally advanced rectal cancer: making sense of the new data. Am Soc Clin Oncol Educ Book. 2022;42:264-277. 10.1200/EDBK_351411 [DOI] [PubMed] [Google Scholar]
- 2. Kotaka M, Shirasu H, Watanabe J, et al. Association of circulating tumor DNA dynamics with clinical outcomes in the adjuvant setting for patients with colorectal cancer from an observational GALAXY study in CIRCULATE-Japan. J Clin Oncol. 2022;40(4_suppl):9-9. 10.1200/jco.2022.40.4_suppl.009 [DOI] [Google Scholar]
- 3. Vidal J, Casadevall D, Bellosillo B, et al. Clinical impact of presurgery circulating tumor DNA after total neoadjuvant treatment in locally advanced rectal cancer: a biomarker study from the GEMCAD 1402 trial. Clin Cancer Res. 2021;27(10):2890-2898. 10.1158/1078-0432.CCR-20-4769 [DOI] [PubMed] [Google Scholar]
- 4. Khakoo S, Carter PD, Brown G, et al. MRI tumor regression grade and circulating tumor DNA as complementary tools to assess response and guide therapy adaptation in rectal cancer. Clin Cancer Res. 2020;26(1):183-192. 10.1158/1078-0432.CCR-19-1996 [DOI] [PubMed] [Google Scholar]
- 5. Zhou J, Wang C, Lin G, et al. Serial circulating tumor DNA in predicting and monitoring the effect of neoadjuvant chemoradiotherapy in patients with rectal cancer: a prospective multicenter study. Clin Cancer Res. 2021;27(1):301-310. 10.1158/1078-0432.CCR-20-2299 [DOI] [PubMed] [Google Scholar]
- 6. Tie J, Cohen JD, Lahouel K, et al. Circulating tumor DNA analysis guiding adjuvant therapy in stage II colon cancer. N Engl J Med. 2022;386(24):2261-2272. 10.1056/NEJMoa2200075 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Maas M, Lambregts DMJ, Nelemans PJ, et al. Assessment of clinical complete response after chemoradiation for rectal cancer with digital rectal examination, endoscopy, and MRI: selection for organ-saving treatment. Ann Surg Oncol. 2015;22(12):3873-3880. 10.1245/s10434-015-4687-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Murahashi S, Akiyoshi T, Sano T, et al. Serial circulating tumour DNA analysis for locally advanced rectal cancer treated with preoperative therapy: prediction of pathological response and postoperative recurrence. Br J Cancer. 2020;123(5):803-810. 10.1038/s41416-020-0941-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Smith JJ, Dasari A, Shi Q, et al. Alliance A022104/NRG-GI010: a randomized phase II trial testing the efficacy of triplet versus doublet chemotherapy to achieve clinical complete response in patients with locally advanced rectal cancer—The Janus Rectal Cancer trial. J Clin Oncol. 2023;41(16_suppl):TPS3640-TPS3640. 10.1200/jco.2023.41.16_suppl.tps3640 [DOI] [Google Scholar]
- 10. GEMCAD-REVEAL STUDY - circulating tumor DNA as a predictor of relapse in patients with locally advanced rectal cancer (REVEAL). Published online June 15, 2023. Accessed June 25, 2023. https://clinicaltrials.gov/study/NCT05674422
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data used to create this article will be shared on reasonable request to the corresponding author.