- Email: [email protected]
Tumor-related leukocytosis is associated with poor radiation response and clinical outcome in uterine cervical cancer patients
Annals of Oncology Advance Access published August 8, 2016 1
Tumor-related leukocytosis is associated with poor radiation response and clinical outcome in uterine cervical cancer patients
Y. Cho1, K.H. Kim1, H.I. Yoon1, G.E. Kim1, Y.B. Kim1,2
1
Department of Radiation Oncology, Yonsei Cancer Center, Yonsei University College of Medicine,
Seoul, Republic of Korea Yonsei Song-Dang Institute for Cancer Research, Yonsei University College of Medicine, Seoul,
Republic of Korea
Corresponding author: Dr. Yong Bae Kim, Department of Radiation Oncology, Yonsei Cancer Center, Yonsei University College of Medicine, 50-1 Yonsei-ro, Seodaemun-gu, Seoul 03722, Republic of Korea. Tel: (+82) 2-2228-8095; Fax: (+82) 2-2227-7823; E-mail: [email protected]
© The Author 2016. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
2
2
Abstract (300) Background: To evaluate response to radiation and clinical outcome of uterine cervical cancer patients with tumor-related leukocytosis (TRL) at initial diagnosis and during definitive radiotherapy. Patients and Methods: We retrospectively analyzed 2,456 patients with stage IA–IVA uterine cervical cancer who received definitive radiotherapy with (37.4%) or without (62.6%) platinum-based chemotherapy between 1986 and 2012. TRL was defined as two or more occurrences of leukocytosis over 9,000/μL at the time of diagnosis and during the course of treatment. Locoregional failure-free
Results: The median age of all patients was 55 years and the median follow-up time was 65.1 months. TRL was observed in 398 patients (16%) at initial diagnosis; TRL (+) patients were younger and had larger tumors, advanced stage, and more frequent lymph node metastases (all p<0.05). TRL (+) patients showed a significantly lower rate of complete remission than TRL (-) patients (89.9% vs. 96.3% respectively, p=0.042). Ten-year LFFS and OS for all patients were 84% and 78%, respectively. LFFS and OS were significantly lower in TRL (+) patients than TRL (-) patients (10-year LFFS: 69% vs. 87% respectively, p<0.001; 10-year OS: 63% vs. 81% respectively p<0.001). After propensity score matching, LFFS and OS rates in TRL (+) patients remained significantly lower than for TRL (-) patients; this significant difference was also observed on multivariate analysis. Twenty-six percent of patients with locoregional failure (n=345) were TRL (+), and had significantly poorer median OS (6 vs. 12 months, p=0.001). Conclusion: This study reveals the aggressive nature of cervical cancer with TRL, and its poor response to radiation therapy. Given the unfavorable prognosis and higher probability of treatment failure, optimal diagnostic and therapeutic approaches and careful monitoring for early detection of recurrence should be considered for these patients.
Keywords: cervical cancer; leukocytosis; neutrophil/lymphocyte ratio; radiation response
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
survival (LFFS) and overall survival (OS) were compared between patients with or without TRL.
3
Key Message: In this large population study of uterine cervical cancer, patients with tumor-related leukocytosis had unfavorable features, and showed poor response to radiotherapy and as well as decreased overall survival. A high neutrophil/lymphocyte ratio is also associated with poor treatment outcome. Therefore, alternative diagnostic and treatment strategies are required for these patients.
Enhanced screening programs and improved socioeconomic conditions have contributed to decreasing the incidence and mortality rates of invasive uterine cervical cancer worldwide. However, the incidence of cervical cancer in Korea remains higher than in the USA and European countries. Therefore, Korean radiation oncologists have more experience treating cervical cancer patients starting early in their careers, and are particularly adept at intracavitary brachytherapy (ICBT). According to the Korea Central Cancer Registry report of 2015, treatment outcomes for uterine cervical cancer in Korea are more favorable than those in the USA and Canada, presumably because of greater clinical experience [1-3]. Nevertheless, patients still suffer from recurrent or uncontrolled disease, and most treatment failures can be explained by conventional prognostic factors including advanced International Federation of Gynecology and Obstetrics (FIGO) stage, larger tumor size, and lymph node (LN) involvement [4]. Failures often occur in some patients who have few poor prognostic factors; therefore, better identification of prognostic criteria that can identify high-risk patients who are resistant to current treatment is required. A number of cancers, including lung, urothelial carcinoma, and sarcoma, are linked to systemic inflammation and elevated white blood cell (WBC) count [5]. Recently, tumor-related leukocytosis (TRL) was reported to be associated with a higher rate of recurrence and dismal prognosis in certain solid cancers [6]. A 2014 Japanese study concluded that TRL was observed in 10– 15% of uterine cervical cancer patients and was associated with resistance to radiotherapy [7]. Additionally, it has been reported that the neutrophil/lymphocyte ratio (NLR) may be related to
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Introduction
4
prognosis in many types of cancer [8, 9]; however, the clinical implication of leukocytosis associated with high NLR has not yet been elucidated. In this study, we evaluated the prognostic value of TRL with high NLR in a large population of cervical cancer patients treated with definitive radiotherapy, and investigated treatment strategies for TRL (+) cervical cancer.
Materials and methods
The study included 2,456 consecutive patients with pathologically confirmed uterine cervical carcinoma, without synchronous distant metastasis, who underwent definitive radiotherapy using external beam radiotherapy (EBRT) and high dose rate intracavitary brachytherapy (ICBT) between 1986 and 2012 (Supplemental Fig. 1). All patients were clinically staged according to the FIGO staging criteria. The routine procedure for staging included a detailed review of patient history, general physical examination, pelvic examination (including bimanual pelvic and rectal examinations), laboratory tests (complete blood-cell count, serum chemistries, and urinalysis), colposcopy, standard chest radiography, intravenous pyelography, cystoscopy, and sigmoidoscopy. Computed tomography (CT) or magnetic resonance image (MRI) were also performed to evaluate the extent of primary tumor and LN involvement. Two hundred and twenty-seven patients (9.2%) underwent 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) since 2004 for evaluating distant metastasis. Patients with synchronous distant metastasis and other primary cancer except thyroid papillary carcinoma were excluded from this study. This study was approved by Institutional Review Board (protocol number: 4-2016-0041). All data were collected by analyzing medical reports and information in our institutional medical records.
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Patient characteristics
5
Treatment profiles EBRT was performed in all patients; those without LN involvement received EBRT to the entire pelvis. In cases with paraaortic LN involvement, extended field radiotherapy to T11/12 or T12/L1 was used. In patients with pelvic LN involvement near the aortic bifurcation, a semi-extended field with the superior border extending to the upper margin of L2 was used [10]. EBRT was delivered at a total dose of 45 Gy, 1.8 Gy per fraction. A weekly pelvic examination was performed to evaluate treatment response, during which we assessed the optimal ICBT applicator geometry within the
prescribed to point A, a point located 2 cm superior and 2 cm lateral to the cervical os, along the plane perpendicular to the intrauterine tandem, the device should pass though the cervical os and reach the uterine fundus. Hence, midline block (MLB) was generally applied at 27–36 Gy depending on the tumor regression. MLB and ICBT procedures were performed as described previously [11]. At our institution, cervical cancer patients who are at high risk of recurrence, including those with advanced FIGO stage disease, bulky tumors, regional lymph node metastasis, and adverse histology, have been undergoing chemotherapy combined with radiotherapy since long before the announcement of the National Cancer Institute consensus statement in 1999. While both induction and concurrent chemotherapy were administered to such patients before 2000, most patients received concurrent chemoradiotherapy (CCRT) after 2000. We have previously reported the details of chemotherapy treatment and subsequent outcomes at our institution [12, 13].
Definitions of TRL and NLR At the study’s commencement (baseline) and during the course of treatment, blood counts of all patients were tested weekly. TRL was defined as at least two separate occasions of leukocytosis exceeding 9,000/µL without any evidence of infection [14]. The NLR was defined as the absolute neutrophil count divided by the absolute lymphocyte count.
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
vaginal cavity that would generate a pear-shaped dose distribution. As the radiation dose is generally
6
Follow-up and clinical outcomes Follow-up examinations were performed every 3 months for the first 2 years, every 6 months for the next 3 years, and then once yearly thereafter. During routine follow up, gynecological pelvic examinations with or without cytology and imaging studies including CT, MRI, or PET-CT were performed. When local recurrence was suspected, biopsy of the lesion was recommended. Recurrences involving the cervix, vagina, or parametrial tissue were classified as local, whereas LN failures within the radiotherapy field were defined as regional recurrence. Recurrence outside the
Statistical Analysis Categorical data were analyzed using Fisher’s exact test or χ2 analyses, and continuous data were compared between groups using the Mann-Whitney U test. The Kaplan-Meier method and logrank test were used to estimate and compare rates of locoregional failure-free survival (LFFS) and overall survival (OS). LFFS and OS rates were measured from the date of treatment commencement to the date of local recurrence and the date of death from any cause, respectively. Logistic regression analysis was performed to analyze the relationship between the complete remission (CR) rate and WBC count. Univariate analyses were performed to identify significant factors affecting LFFS and OS. We further conducted multivariate Cox hazard regression analysis on variables with p values <0.05 on univariate analysis. To balance the patient and tumor characteristics between TRL (-) and TRL (+) groups, propensity score matching (PSM) was performed using the Match-it package for R software, version 3.1.0. Propensity scores were estimated by multivariate logistic regression models for the probability of TRL (+) adjusting for age, FIGO stage, histology, tumor size, LN metastasis, and pre-treatment hemoglobin. TRL (+) patients were matched with TRL (-) counterparts according to propensity scores, leading to an even distribution of potential confounding factors in the treatment groups. P<0.05 was considered statistically significant. All analyses were performed using IBM SPSS version 20.0 (SPSS, Chicago, IL).
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
radiotherapy field was defined as distant metastasis.
7
Results Patient and treatment characteristics Patient and treatment characteristics of 2,456 patients, consisting of 398 (16.2%) patients in the TRL (+) group and 2,058 in the TRL (-) group (83.8%), are presented in Table 1. TRL (+) group patients, who were significantly younger, had poorer prognostic characteristics compared to TRL (-)
treatment hemoglobin. The median NLR of all patients was 2.17; those with high NLR (>2.5) were more frequent in the TRL (+) group (79% vs. 34%, p<0.001). The absolute neutrophil count was also higher in the TRL (+) group. FDG-PET records were available for 225 patients; 182 in the TRL (-) group (8.8%) and 43 patients in TRL (+) group (10.8%). Among these patients, bone marrow hypermetabolism was observed in 32 TRL (-) patients and 27 TRL (+) patients (17.6% vs. 62.8%, respectively, p<0.001) (Fig. 1). Concurrent or induction chemotherapy was administered to 918 patients (37.4%); concurrent chemotherapy to 585 patients (23.8%), and induction chemotherapy to 333 patients (13.6%). Most received platinum-based chemotherapy. Concurrent chemotherapy included a combination of carboplatin or cisplatin with 5-fluorouracil performed in the first, fourth, and seventh weeks of radiotherapy (n=317, 54.2%), or weekly administration of cisplatin during radiotherapy (n=244, 41.7%). For patients who received induction chemotherapy, platinum-based chemotherapy was administered every 3 weeks in 155 patients (46.5%), weekly cisplatin in 13 patients (4%), and other treatments in 165 patients (49.5%). The median dose to point A was greater (80.3 vs. 76.9 Gy, p<0.001) and the elapsed days of treatment were longer in the TRL (+) group (76 vs. 71 days, p<0.001). Nevertheless, patients in the TRL (+) group had a lower CR rate (89.9% vs. 95.2%, p<0.001).
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
patients, including advanced clinical stage, bulky tumor, more LN metastases, and lower pre-
8
Patterns of failures The sites of first tumor recurrence are presented in Table 2. The overall recurrence rate was significantly higher in the TRL (+) group than in the TRL (-) group (40.5% vs. 19% respectively, p<0.001). Locoregional and distant recurrences were more frequent in the TRL (+) group (28.1% vs. 11.4%, p<0.001; and 17.8% vs. 11.2%, p<0.001; respectively). Similarly, recurrences of any type were more frequent in the high NLR (>2.5) group than those in the low NLR (≤2.5) group. Locoregional and distant recurrence rates were also higher in the
Survival and prognostic analysis The median follow up time was 65.1 (range, 0.7–347.8) months. The 10-year LFFS and OS rate for all patients were 84% and 78%, respectively. As shown in Table 3, multivariate analyses show that patients’ age, clinical stage, histologic type, tumor size, and TRL all correlated with poor LFFS; the NLR did not provide prognostic information regarding locoregional failure. Advanced clinical stage, non-squamous cell carcinoma, larger tumor size, LN metastasis, and leukocytosis were identified as significant prognostic factors for OS. Additionally, high NLR was associated with poor OS on multivariate analysis (hazard ratio [HR] 1.313, 95% confidence interval [CI] 1.05–1.618, p=0.016). LFFS and OS rate were significantly lower in the TRL (+) group than in the TRL (-) groups (10-year LFFS: 69% vs. 87% respectively, p<0.001; 10-year OS: 63% vs. 81% respectively, p<0.001) (Fig. 2A). All patients were divided in four groups according to their TRL and NLR statuses. Poorer LFFS was observed in more patients with higher NLR (>2.5) than those with low NLR (≤2.5) in the TRL (+) group; NLR did not influence LFFS in the TRL (-) group. Patients with high NLR showed OS inferior to low NLR patients in both the TRL (-) and TRL (+) groups (Fig. 2B). After PSM, patient and tumor characteristics were well matched between the two groups (Supplemental Table 1); the LFFS and OS of the TRL (+) group remained poorer compared to the TRL (-) group (10-year LFFS: 69.1% vs. 80.1% respectively, p<0.001; 10-year OS: 63.6% vs. 72.1%
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
high NLR group (17.4% vs. 11.3%, p<0.001; and 14.8% vs. 11.1%, p<0.001; respectively).
9
respectively, p=0.005) (Supplemental Fig. 2).
Outcome according to treatment scheme Patients with CCRT showed more favorable LFFS than those with induction chemotherapy followed by RT (induction CRT) or RT alone (10-year LFFS for CCRT vs. induction CRT vs. RT alone: 88.2% vs. 74.8% vs. 84.0%, respectively, p<0.001). OS was significantly lower in patients with induction CRT (10-year OS for CCRT vs. induction CRT vs. RT alone: 78.6% vs. 71.6% vs. 79.8%,
(Table 3). When we compared the outcomes according to TRL status in each treatment group, patients with TRL showed inferior LFFS and OS across all treatment groups. The chemotherapy regimen (weekly cisplatin vs. every 3 weeks) did not affect the LFFS or OS (10-year LFFS: 86.2% vs. 81.4, respectively, p=0.140; 10-year OS: 71.7% vs. 74.7% respectively, p=0.823).
TRL at the time of first recurrence Among patients who experienced locoregional recurrence (n=232), 60 (25.8%) showed TRL at the time of first recurrence, which was a higher proportion of TRL positivity than at initial diagnosis. Survival after recurrence in these patients was inferior (median OS 6 vs. 12 months, p<0.001) (Supplemental Fig. 3).
Discussion Radiotherapy is a major treatment modality for uterine cervical cancer, and tumor response to radiation is important for determining success or failure of treatment; i.e., CR is the most significant index of radiotherapy effectiveness [15]. Although a high CR rate is generally associated with a higher physical radiation dose, the appropriate dose should weigh both radiation-related toxicities and tumor control. Aside from radiation dose, there might be unknown biological differences between CR responders and non-responders; moreover, the time to achieve CR might be
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
respectively, p<0.001). CCRT was correlated with improved LFFS and OS on multivariate analysis
10
associated with treatment outcomes. To that end, TRL has recently been implicated in response to radiotherapy [7, 16]. Studies have long revealed that inflammation is involved in carcinogenesis and tumor progression [5]. Inflammatory reactions have sometimes been regarded as pre-cancerous lesions in several organs including head and neck, lung, and uterine cervix; the inflammatory microenvironment can nourish tumor development and progression. TRL and high NLR are systemic responses of the inflammatory process, and appear to be associated with poor prognosis in certain types of cancer.
increased levels of several inflammatory cytokines contribute to creating a microenvironment that promotes carcinogenesis and tumor progression [17]. Several growth factors such as EGF, VEGF, and TGF-α also contribute to creating microenvironments that promote angiogenesis and tumor proliferation [18]. Moreover, myeloid-derived suppressor cells (MDSCs) decrease the activity of antitumor cytotoxic lymphocytes, accelerating tumor growth [19]. In 2014, Mabuchi et al. showed that uterine cervical cancer patients with elevated WBC counts were younger and had larger tumors and advanced stages. Moreover, TRL was associated with poor OS and resistance to radiotherapy, and TRL (+) patients showed significantly increased expression of granulocyte colony-stimulating factor (G-CSF) [7]. G-CSF-derived tumor increases MDSC in the bone marrow, blood stream, and spleen; this promotes a favorable microenvironment for tumor growth. Although the connection between MDSC and tumor growth has not yet been identified, other studies revealed that elevated levels of MDSC are associated with tumor proliferation. Thus, employing anti-MDSC agents could improve tumor response to radiation. Additionally, given that the spleen supplies MDSCs (hence the designation “splenic reservoir”), suppression of MDSC levels via splenectomy impairs G-CSF-derived tumor progression [7, 20, 21]. This study confirms the clinical significance of TRL and NLR in a large uterine cervical cancer patient population. Unfavorable prognostic features such as younger age, advanced FIGO stage, larger tumor size, and more LN metastases were more frequently accompanied by TRL. TRL (+)
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
However, the tumorigenic mechanism has not been clearly elucidated, although it appears likely that
11
patients also showed poor responses to radiotherapy and lower CR rates despite receiving higher doses, resulting in decreased OS; TRL remained a strong prognostic and predictive factor after performing PSM. Additionally, bone marrow hypermetabolism on FDG-PET occurred more frequently in TRL (+) patients, supporting the hypothesis that radioresistance is associated with increased levels of G-CSF-induced MDSC. Of note, a high NLR strongly correlated with poorer OS. In a previous study, patients with high NLRs showed low responses to local RT and granulocyte-macrophage colony-stimulating factor.
results are consistent with the notion that a high NLR is associated with the poor systemic disease control by suppressing the cytolytic activity of activated effector T cells as well as peritumoral infiltration of immunosuppressive macrophages [19]. Our findings strongly suggest that novel treatment strategies are required for certain patients. Namely, patients with TRL showing leukocytosis or neutrophilia at baseline, as well as those with bone marrow hyperactivity on PET, an enlarged spleen, or elevated WBC counts should be closely monitored during treatment, including their response to radiotherapy. Intensified radiotherapy (including simultaneous integrated boost or dose escalation) may improve local control for these patients. Moreover, systemic MDSC suppression treatment may be beneficial. A drawback of this study is that it was performed retrospectively over a long period of time. Furthermore, the selection of treatment modalities and regimens were heterogeneous throughout this period. Therefore, our findings should be interpreted with some caution. Conclusions Uterine cervical cancer with TRL is aggressive in nature and shows a poor response to radiation therapy. Our large population study validated the notion that both TRL and high NLR are associated with poorer outcomes. While further studies are required to investigate the biological mechanisms of TRL-promoted aggressiveness, our evidence clearly indicates that alternative
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Additionally, such patients did not benefit from the abscopal effect [22]. Along the same lines, our
12
diagnostic and treatment strategies are warranted for such patients.
ACKNOWLEDGEMENTS: None FUNDING: There was no research funding for this study. DISCLOSURE: The authors have declared no conflicts of interest relevant to this manuscript.
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
13
References 1.
Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer J Clin 2015; 65: 5-29.
2.
Oh CM, Won YJ, Jung KW et al. Cancer Statistics in Korea: Incidence, Mortality, Survival,
and Prevalence in 2013. Cancer Res Treat 2016; 48: 436-450. 3.
Annual report of cancer statistics in Korea in 2013. In. Korea Central Cancer Registry 2015.
4.
Kapp DS, Fischer D, Gutierrez E et al. Pretreatment prognostic factors in carcinoma of the
uterine cervix: a multivariable analysis of the effect of age, stage, histology and blood counts on survival. Int J Radiat Oncol Biol Phys 1983; 9: 445-455. 5.
Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature 2008; 454:
436-444. 6.
Granger JM, Kontoyiannis DP. Etiology and outcome of extreme leukocytosis in 758
nonhematologic cancer patients: a retrospective, single-institution study. Cancer 2009; 115: 39193923. 7.
Mabuchi S, Matsumoto Y, Kawano M et al. Uterine cervical cancer displaying tumor-
related leukocytosis: a distinct clinical entity with radioresistant feature. J Natl Cancer Inst 2014; 106. 8.
Lee YY, Choi CH, Kim HJ et al. Pretreatment neutrophil:lymphocyte ratio as a prognostic
factor in cervical carcinoma. Anticancer Res 2012; 32: 1555-1561. 9.
Teramukai S, Kitano T, Kishida Y et al. Pretreatment neutrophil count as an independent
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
http://ncc.re.kr/main.ncc?uri=english/sub04_Statistics
14
prognostic factor in advanced non-small-cell lung cancer: an analysis of Japan Multinational Trial Organisation LC00-03. Eur J Cancer 2009; 45: 1950-1958. 10.
Choi J, Yoon HI, Lee J et al. Optimal Extent of Prophylactic Irradiation of Paraaortic Lymph
Nodes in Patients with Uterine Cervical Cancer. PLoS One 2015; 10: e0145158. 11.
Kim KH, Kim S, Kim GE et al. De-escalation of the cumulative central radiation dose
outcome in patients with uterine cervical cancer. Gynecol Oncol 2015; 139: 439-446. 12.
Kim YB, Cho JH, Keum KC et al. Concurrent chemoradiotherapy followed by adjuvant
chemotherapy in uterine cervical cancer patients with high-risk factors. Gynecol Oncol 2007; 104: 58-63. 13.
Park TK, Choi DH, Kim SN et al. Role of induction chemotherapy in invasive cervical
cancer. Gynecol Oncol 1991; 41: 107-112. 14.
Mabuchi S, Matsumoto Y, Hamasaki T et al. Elevated white blood cell count at the time of
recurrence diagnosis is an indicator of short survival in patients with recurrent cervical cancer. Int J Gynecol Cancer 2012; 22: 1545-1551. 15.
Kim YB, Lee IJ, Kim SY et al. Tumor heterogeneity of FIGO stage III carcinoma of the
uterine cervix. Int J Radiat Oncol Biol Phys 2009; 75: 1323-1328. 16.
Mabuchi S, Matsumoto Y, Isohashi F et al. Pretreatment leukocytosis is an indicator of
poor prognosis in patients with cervical cancer. Gynecol Oncol 2011; 122: 25-32. 17.
Schetter AJ, Heegaard NH, Harris CC. Inflammation and cancer: interweaving microRNA,
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
according to the tumor response can reduce rectal toxicity without compromising the treatment
15
free radical, cytokine and p53 pathways. Carcinogenesis 2010; 31: 37-49. 18.
Gupta RA, Tejada LV, Tong BJ et al. Cyclooxygenase-1 is overexpressed and promotes
angiogenic growth factor production in ovarian cancer. Cancer Res 2003; 63: 906-911. 19.
Templeton AJ, McNamara MG, Seruga B et al. Prognostic role of neutrophil-to-
lymphocyte ratio in solid tumors: a systematic review and meta-analysis. J Natl Cancer Inst 2014;
20.
Abrams SI, Waight JD. Identification of a G-CSF-Granulocytic MDSC axis that promotes
tumor progression. Oncoimmunology 2012; 1: 550-551. 21.
Galdiero MR, Garlanda C, Jaillon S et al. Tumor associated macrophages and neutrophils
in tumor progression. J Cell Physiol 2013; 228: 1404-1412. 22.
Golden EB, Chhabra A, Chachoua A et al. Local radiotherapy and granulocyte-
macrophage colony-stimulating factor to generate abscopal responses in patients with metastatic solid tumours: a proof-of-principle trial. Lancet Oncol 2015; 16: 795-803.
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
106: dju124.
16
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
17
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Table 1. Patient and treatment characteristics All patients Characteristic
No.
TRL (-) group (%)
No.
TRL (+) group (%)
No.
n=2456
n=2058
n=398
56 (21-87)
56 (21-87)
53 (25-86)
(%)
P
Age Median (range)
<0.001
Clinical stage 494
20.1
465
22.6
29
7.3
IIA-IIB
1530
62.3
1288
62.6
242
60.8
IIIA-IIIB
394
16
280
13.6
114
28.6
IVA
38
1.5
25
1.2
13
3.3
Histologic subtype SCCa
2263
92.1
1988
92.2
365
91.7
non-SCCa
193
7.9
160
7.8
33
8.3
Tumor size (mm) Median (range)
40 (2-111)
40 (2-105)
50 (6-111)
<0.001
0.726
<0.001
<40
1046
42.6
959
46.6
87
21.9
40-50
570
23.2
490
23.8
80
20.1
≥51
840
34.2
609
29.6
231
58
Yes
516
21
394
19.1
122
30.7
No
1940
79
1664
80.9
276
69.3
≤10
207
8.4
145
7
62
15.6
>10
2249
91.6
1913
93
336
84.4
LN metastasis <0.001
Pre Tx Hb (mg/dl) <0.001
NLR Median (range)
2.17 (0.2-48)
2 (0.2-31.7)
4.3 (0.7-48)
<0.001
≤2.5 >2.5
1341
58.8
1260
66.1
81
21.5
940
41.2
645
33.9
295
78.5
Median (range)
4,088
3,762
8,024
(632-3,5512)
(632-3,5512)
(842-3,1574)
BM hypermetabolism Yes
59
32
17.6
27
62.8
No
166
150
82.4
16
37.2
ANC <0.001
<0.001
Treatment RT alone
1538
62.6
1322
64.2
216
54.3
Induction chemo +RT
333
13.6
266
12.9
67
16.8
CCRT
585
23.8
470
22.8
115
28.9
<0.001
Point A dose* Median
77.6
76.9
80.3
(range)
(51.7-122.2)
(51.7-122.2)
(59.1-101.6)
Duration of RT Median (range)
72 (36-181)
71 (36-171)
76 (45-181)
<0.001
<0.001
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
IA-IB
RT response CR
2339
95.2
1981
96.3
358
89.9
Non CR
117
4.8
77
3.7
40
10.1
<0.001
Abbreviation: SCCa; squamous cell carcinoma, LN; lymph node, Pre Tx Hb; pre-treatment hemoglobin, NLR: neutrophil/lymphocyte ratio, ANC; absolute neutrophil count, BM; bone marrow, RT; radiotherapy, CCRT; concurrent chemoradiotherapy CR; complete remission *EQD2, α/β=10
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Table 2. Patterns of failure according to tumor-related leukocytosis (TRL) and neutriophil/lymphocyte ratio (NLR) TRL (-) group
TRL (+) group
Low NLR
n = 2058
n = 398
n = 1341
No.
%
No.
%
P
No.
%
391
19
161
40.5
<0.001
256
19.1
Locoregional
234
11.4
112
28.1
<0.001
152
11.3
Distant
230
11.2
71
17.8
<0.001
149
11.1
Any recurrence
High NLR n = 940
No.
%
P
260
27.7
<0.001
164
17.4
<0.001
139
14.8
0.009
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Table 3. Univariate and multivariate analysis for locoregional failure-free survival (LFFS) and overall survival (OS)
Variable
HR
LFFS
LFFS
OS
Univariate analysis
Multivariate analysis
Univariate analysis
95%CI
P
HR
95%CI
P
HR
95%CI
Age <56
1
≥56
0.688
1 0.055-0.853
0.001
0.72
1 0.661-0.975
0.029
0.879
0.734-1.053
Clinical stage IA-IB
1
1
1
OS Multivariate analysis P
HR
95%CI
P
0.161 1
IIA-IIB
2.787
1.865-4.164
<0.001
2.331
1.524-3.567
<0.001
2.099
1.566-2.831
<0.001
1.773
1.285-2.448
<0.001
IIIA-IIIB
6.963
4.555-10.643
<0.001
4.166
2.58-6.726
<0.001
5.85
4.24-8.071
<0.001
3.533
2.442-5.110
<0.001
IVA
9.731
4.821-19.644
<0.001
6.46
3.0.71-13.59
<0.001
10.548
6.275-17.73
<0.001
6.604
3.785-11.523
<0.001
SCC
1
1.318-2.351
<0.001
1.225-1.865
<0.001
1.618-2.45
<0.001
0.843-1.537
0.398
1.039-2.316
0.032
Histology Non-SCC Tumor size (mm) <40 ≥40
1.568
1 1.124-2.189
0.008
1 1.791
1.782
1 1.264-2.513
0.001
1 1.308-2.452
<0.001
1.619
1.23-1.977
<0.001
1.265
1.504
1.133-1.966
1 1.258-2.084
<0.001
2.546
0.972-1.647
0.081
2.363
0.005
1 1.76 1
2.219-3.044
<0.001
1.512
1.961-2.846
<0.001
1.991
LN metastasis No
1
Yes
1.56
Pre Tx Hb (mg/dl) <10
1
1
>10
0.654
No
1
Yes
2.944
1
1 0.469-0.913
0.018
1.189
1
1 0.816-1.731
0.367
0.563
1 0.43-0.736
<0.001
1.138
Leukocytosis
NLR
1 2.35-3.689
<0.001
2.103
1 1.615-2.737
<0.001
2.308
1 1.885-2.826
<0.001
1.552
1
>2.5
1.741
<0.001 1.396-2.171
1 1.065
1 0.830-1.365
0.621
1.897
1.576-2.284
ANC <6,000
1
≥6,000
2.808
1 2.266-3.542
<0.001
1.264
1 0.780-2.047
0.342
2.217
1.806-2.722
Treatment RT alone
1
1
1
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
≤2.5
<0.001
<0.001
1 1.305
1.051-1.62
0.016
0.626-1.443
0.81
1 0.95 1
Induction CRT
1.739
1337-2.261
<000.1
1.280
0.963-1.701
0.089
1.538
1.214-1.949
<0.001
1.023
0.794-1.318
0.861
CCRT
0.745
0.561-0.989
0.042
0.611
0.448-0.832
0.002
1.016
0.817-1.264
0.887
0.762
0.6-0.969
0.027
Abbreviations: SCCa; squamous cell carcinoma, LN; lymph node, Pre Tx Hb; pre-treatment hemoglobin, NLR: neutrophil/lymphocyte ratio, ANC; absolute neutrophil count, Induction CRT; induction chemotherapy followed by radiotherapy, CCRT: concurrent chemoradiotherapy
Tumor-related leukocytosis is associated with poor radiation response and clinical outcome in uterine cervical cancer patients
Y. Cho1, K.H. Kim1, H.I. Yoon1, G.E. Kim1, Y.B. Kim1,2
1
Department of Radiation Oncology, Yonsei Cancer Center, Yonsei University College of Medicine,
Seoul, Republic of Korea Yonsei Song-Dang Institute for Cancer Research, Yonsei University College of Medicine, Seoul,
Republic of Korea
Corresponding author: Dr. Yong Bae Kim, Department of Radiation Oncology, Yonsei Cancer Center, Yonsei University College of Medicine, 50-1 Yonsei-ro, Seodaemun-gu, Seoul 03722, Republic of Korea. Tel: (+82) 2-2228-8095; Fax: (+82) 2-2227-7823; E-mail: [email protected]
© The Author 2016. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
2
2
Abstract (300) Background: To evaluate response to radiation and clinical outcome of uterine cervical cancer patients with tumor-related leukocytosis (TRL) at initial diagnosis and during definitive radiotherapy. Patients and Methods: We retrospectively analyzed 2,456 patients with stage IA–IVA uterine cervical cancer who received definitive radiotherapy with (37.4%) or without (62.6%) platinum-based chemotherapy between 1986 and 2012. TRL was defined as two or more occurrences of leukocytosis over 9,000/μL at the time of diagnosis and during the course of treatment. Locoregional failure-free
Results: The median age of all patients was 55 years and the median follow-up time was 65.1 months. TRL was observed in 398 patients (16%) at initial diagnosis; TRL (+) patients were younger and had larger tumors, advanced stage, and more frequent lymph node metastases (all p<0.05). TRL (+) patients showed a significantly lower rate of complete remission than TRL (-) patients (89.9% vs. 96.3% respectively, p=0.042). Ten-year LFFS and OS for all patients were 84% and 78%, respectively. LFFS and OS were significantly lower in TRL (+) patients than TRL (-) patients (10-year LFFS: 69% vs. 87% respectively, p<0.001; 10-year OS: 63% vs. 81% respectively p<0.001). After propensity score matching, LFFS and OS rates in TRL (+) patients remained significantly lower than for TRL (-) patients; this significant difference was also observed on multivariate analysis. Twenty-six percent of patients with locoregional failure (n=345) were TRL (+), and had significantly poorer median OS (6 vs. 12 months, p=0.001). Conclusion: This study reveals the aggressive nature of cervical cancer with TRL, and its poor response to radiation therapy. Given the unfavorable prognosis and higher probability of treatment failure, optimal diagnostic and therapeutic approaches and careful monitoring for early detection of recurrence should be considered for these patients.
Keywords: cervical cancer; leukocytosis; neutrophil/lymphocyte ratio; radiation response
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
survival (LFFS) and overall survival (OS) were compared between patients with or without TRL.
3
Key Message: In this large population study of uterine cervical cancer, patients with tumor-related leukocytosis had unfavorable features, and showed poor response to radiotherapy and as well as decreased overall survival. A high neutrophil/lymphocyte ratio is also associated with poor treatment outcome. Therefore, alternative diagnostic and treatment strategies are required for these patients.
Enhanced screening programs and improved socioeconomic conditions have contributed to decreasing the incidence and mortality rates of invasive uterine cervical cancer worldwide. However, the incidence of cervical cancer in Korea remains higher than in the USA and European countries. Therefore, Korean radiation oncologists have more experience treating cervical cancer patients starting early in their careers, and are particularly adept at intracavitary brachytherapy (ICBT). According to the Korea Central Cancer Registry report of 2015, treatment outcomes for uterine cervical cancer in Korea are more favorable than those in the USA and Canada, presumably because of greater clinical experience [1-3]. Nevertheless, patients still suffer from recurrent or uncontrolled disease, and most treatment failures can be explained by conventional prognostic factors including advanced International Federation of Gynecology and Obstetrics (FIGO) stage, larger tumor size, and lymph node (LN) involvement [4]. Failures often occur in some patients who have few poor prognostic factors; therefore, better identification of prognostic criteria that can identify high-risk patients who are resistant to current treatment is required. A number of cancers, including lung, urothelial carcinoma, and sarcoma, are linked to systemic inflammation and elevated white blood cell (WBC) count [5]. Recently, tumor-related leukocytosis (TRL) was reported to be associated with a higher rate of recurrence and dismal prognosis in certain solid cancers [6]. A 2014 Japanese study concluded that TRL was observed in 10– 15% of uterine cervical cancer patients and was associated with resistance to radiotherapy [7]. Additionally, it has been reported that the neutrophil/lymphocyte ratio (NLR) may be related to
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Introduction
4
prognosis in many types of cancer [8, 9]; however, the clinical implication of leukocytosis associated with high NLR has not yet been elucidated. In this study, we evaluated the prognostic value of TRL with high NLR in a large population of cervical cancer patients treated with definitive radiotherapy, and investigated treatment strategies for TRL (+) cervical cancer.
Materials and methods
The study included 2,456 consecutive patients with pathologically confirmed uterine cervical carcinoma, without synchronous distant metastasis, who underwent definitive radiotherapy using external beam radiotherapy (EBRT) and high dose rate intracavitary brachytherapy (ICBT) between 1986 and 2012 (Supplemental Fig. 1). All patients were clinically staged according to the FIGO staging criteria. The routine procedure for staging included a detailed review of patient history, general physical examination, pelvic examination (including bimanual pelvic and rectal examinations), laboratory tests (complete blood-cell count, serum chemistries, and urinalysis), colposcopy, standard chest radiography, intravenous pyelography, cystoscopy, and sigmoidoscopy. Computed tomography (CT) or magnetic resonance image (MRI) were also performed to evaluate the extent of primary tumor and LN involvement. Two hundred and twenty-seven patients (9.2%) underwent 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) since 2004 for evaluating distant metastasis. Patients with synchronous distant metastasis and other primary cancer except thyroid papillary carcinoma were excluded from this study. This study was approved by Institutional Review Board (protocol number: 4-2016-0041). All data were collected by analyzing medical reports and information in our institutional medical records.
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Patient characteristics
5
Treatment profiles EBRT was performed in all patients; those without LN involvement received EBRT to the entire pelvis. In cases with paraaortic LN involvement, extended field radiotherapy to T11/12 or T12/L1 was used. In patients with pelvic LN involvement near the aortic bifurcation, a semi-extended field with the superior border extending to the upper margin of L2 was used [10]. EBRT was delivered at a total dose of 45 Gy, 1.8 Gy per fraction. A weekly pelvic examination was performed to evaluate treatment response, during which we assessed the optimal ICBT applicator geometry within the
prescribed to point A, a point located 2 cm superior and 2 cm lateral to the cervical os, along the plane perpendicular to the intrauterine tandem, the device should pass though the cervical os and reach the uterine fundus. Hence, midline block (MLB) was generally applied at 27–36 Gy depending on the tumor regression. MLB and ICBT procedures were performed as described previously [11]. At our institution, cervical cancer patients who are at high risk of recurrence, including those with advanced FIGO stage disease, bulky tumors, regional lymph node metastasis, and adverse histology, have been undergoing chemotherapy combined with radiotherapy since long before the announcement of the National Cancer Institute consensus statement in 1999. While both induction and concurrent chemotherapy were administered to such patients before 2000, most patients received concurrent chemoradiotherapy (CCRT) after 2000. We have previously reported the details of chemotherapy treatment and subsequent outcomes at our institution [12, 13].
Definitions of TRL and NLR At the study’s commencement (baseline) and during the course of treatment, blood counts of all patients were tested weekly. TRL was defined as at least two separate occasions of leukocytosis exceeding 9,000/µL without any evidence of infection [14]. The NLR was defined as the absolute neutrophil count divided by the absolute lymphocyte count.
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
vaginal cavity that would generate a pear-shaped dose distribution. As the radiation dose is generally
6
Follow-up and clinical outcomes Follow-up examinations were performed every 3 months for the first 2 years, every 6 months for the next 3 years, and then once yearly thereafter. During routine follow up, gynecological pelvic examinations with or without cytology and imaging studies including CT, MRI, or PET-CT were performed. When local recurrence was suspected, biopsy of the lesion was recommended. Recurrences involving the cervix, vagina, or parametrial tissue were classified as local, whereas LN failures within the radiotherapy field were defined as regional recurrence. Recurrence outside the
Statistical Analysis Categorical data were analyzed using Fisher’s exact test or χ2 analyses, and continuous data were compared between groups using the Mann-Whitney U test. The Kaplan-Meier method and logrank test were used to estimate and compare rates of locoregional failure-free survival (LFFS) and overall survival (OS). LFFS and OS rates were measured from the date of treatment commencement to the date of local recurrence and the date of death from any cause, respectively. Logistic regression analysis was performed to analyze the relationship between the complete remission (CR) rate and WBC count. Univariate analyses were performed to identify significant factors affecting LFFS and OS. We further conducted multivariate Cox hazard regression analysis on variables with p values <0.05 on univariate analysis. To balance the patient and tumor characteristics between TRL (-) and TRL (+) groups, propensity score matching (PSM) was performed using the Match-it package for R software, version 3.1.0. Propensity scores were estimated by multivariate logistic regression models for the probability of TRL (+) adjusting for age, FIGO stage, histology, tumor size, LN metastasis, and pre-treatment hemoglobin. TRL (+) patients were matched with TRL (-) counterparts according to propensity scores, leading to an even distribution of potential confounding factors in the treatment groups. P<0.05 was considered statistically significant. All analyses were performed using IBM SPSS version 20.0 (SPSS, Chicago, IL).
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
radiotherapy field was defined as distant metastasis.
7
Results Patient and treatment characteristics Patient and treatment characteristics of 2,456 patients, consisting of 398 (16.2%) patients in the TRL (+) group and 2,058 in the TRL (-) group (83.8%), are presented in Table 1. TRL (+) group patients, who were significantly younger, had poorer prognostic characteristics compared to TRL (-)
treatment hemoglobin. The median NLR of all patients was 2.17; those with high NLR (>2.5) were more frequent in the TRL (+) group (79% vs. 34%, p<0.001). The absolute neutrophil count was also higher in the TRL (+) group. FDG-PET records were available for 225 patients; 182 in the TRL (-) group (8.8%) and 43 patients in TRL (+) group (10.8%). Among these patients, bone marrow hypermetabolism was observed in 32 TRL (-) patients and 27 TRL (+) patients (17.6% vs. 62.8%, respectively, p<0.001) (Fig. 1). Concurrent or induction chemotherapy was administered to 918 patients (37.4%); concurrent chemotherapy to 585 patients (23.8%), and induction chemotherapy to 333 patients (13.6%). Most received platinum-based chemotherapy. Concurrent chemotherapy included a combination of carboplatin or cisplatin with 5-fluorouracil performed in the first, fourth, and seventh weeks of radiotherapy (n=317, 54.2%), or weekly administration of cisplatin during radiotherapy (n=244, 41.7%). For patients who received induction chemotherapy, platinum-based chemotherapy was administered every 3 weeks in 155 patients (46.5%), weekly cisplatin in 13 patients (4%), and other treatments in 165 patients (49.5%). The median dose to point A was greater (80.3 vs. 76.9 Gy, p<0.001) and the elapsed days of treatment were longer in the TRL (+) group (76 vs. 71 days, p<0.001). Nevertheless, patients in the TRL (+) group had a lower CR rate (89.9% vs. 95.2%, p<0.001).
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
patients, including advanced clinical stage, bulky tumor, more LN metastases, and lower pre-
8
Patterns of failures The sites of first tumor recurrence are presented in Table 2. The overall recurrence rate was significantly higher in the TRL (+) group than in the TRL (-) group (40.5% vs. 19% respectively, p<0.001). Locoregional and distant recurrences were more frequent in the TRL (+) group (28.1% vs. 11.4%, p<0.001; and 17.8% vs. 11.2%, p<0.001; respectively). Similarly, recurrences of any type were more frequent in the high NLR (>2.5) group than those in the low NLR (≤2.5) group. Locoregional and distant recurrence rates were also higher in the
Survival and prognostic analysis The median follow up time was 65.1 (range, 0.7–347.8) months. The 10-year LFFS and OS rate for all patients were 84% and 78%, respectively. As shown in Table 3, multivariate analyses show that patients’ age, clinical stage, histologic type, tumor size, and TRL all correlated with poor LFFS; the NLR did not provide prognostic information regarding locoregional failure. Advanced clinical stage, non-squamous cell carcinoma, larger tumor size, LN metastasis, and leukocytosis were identified as significant prognostic factors for OS. Additionally, high NLR was associated with poor OS on multivariate analysis (hazard ratio [HR] 1.313, 95% confidence interval [CI] 1.05–1.618, p=0.016). LFFS and OS rate were significantly lower in the TRL (+) group than in the TRL (-) groups (10-year LFFS: 69% vs. 87% respectively, p<0.001; 10-year OS: 63% vs. 81% respectively, p<0.001) (Fig. 2A). All patients were divided in four groups according to their TRL and NLR statuses. Poorer LFFS was observed in more patients with higher NLR (>2.5) than those with low NLR (≤2.5) in the TRL (+) group; NLR did not influence LFFS in the TRL (-) group. Patients with high NLR showed OS inferior to low NLR patients in both the TRL (-) and TRL (+) groups (Fig. 2B). After PSM, patient and tumor characteristics were well matched between the two groups (Supplemental Table 1); the LFFS and OS of the TRL (+) group remained poorer compared to the TRL (-) group (10-year LFFS: 69.1% vs. 80.1% respectively, p<0.001; 10-year OS: 63.6% vs. 72.1%
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
high NLR group (17.4% vs. 11.3%, p<0.001; and 14.8% vs. 11.1%, p<0.001; respectively).
9
respectively, p=0.005) (Supplemental Fig. 2).
Outcome according to treatment scheme Patients with CCRT showed more favorable LFFS than those with induction chemotherapy followed by RT (induction CRT) or RT alone (10-year LFFS for CCRT vs. induction CRT vs. RT alone: 88.2% vs. 74.8% vs. 84.0%, respectively, p<0.001). OS was significantly lower in patients with induction CRT (10-year OS for CCRT vs. induction CRT vs. RT alone: 78.6% vs. 71.6% vs. 79.8%,
(Table 3). When we compared the outcomes according to TRL status in each treatment group, patients with TRL showed inferior LFFS and OS across all treatment groups. The chemotherapy regimen (weekly cisplatin vs. every 3 weeks) did not affect the LFFS or OS (10-year LFFS: 86.2% vs. 81.4, respectively, p=0.140; 10-year OS: 71.7% vs. 74.7% respectively, p=0.823).
TRL at the time of first recurrence Among patients who experienced locoregional recurrence (n=232), 60 (25.8%) showed TRL at the time of first recurrence, which was a higher proportion of TRL positivity than at initial diagnosis. Survival after recurrence in these patients was inferior (median OS 6 vs. 12 months, p<0.001) (Supplemental Fig. 3).
Discussion Radiotherapy is a major treatment modality for uterine cervical cancer, and tumor response to radiation is important for determining success or failure of treatment; i.e., CR is the most significant index of radiotherapy effectiveness [15]. Although a high CR rate is generally associated with a higher physical radiation dose, the appropriate dose should weigh both radiation-related toxicities and tumor control. Aside from radiation dose, there might be unknown biological differences between CR responders and non-responders; moreover, the time to achieve CR might be
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
respectively, p<0.001). CCRT was correlated with improved LFFS and OS on multivariate analysis
10
associated with treatment outcomes. To that end, TRL has recently been implicated in response to radiotherapy [7, 16]. Studies have long revealed that inflammation is involved in carcinogenesis and tumor progression [5]. Inflammatory reactions have sometimes been regarded as pre-cancerous lesions in several organs including head and neck, lung, and uterine cervix; the inflammatory microenvironment can nourish tumor development and progression. TRL and high NLR are systemic responses of the inflammatory process, and appear to be associated with poor prognosis in certain types of cancer.
increased levels of several inflammatory cytokines contribute to creating a microenvironment that promotes carcinogenesis and tumor progression [17]. Several growth factors such as EGF, VEGF, and TGF-α also contribute to creating microenvironments that promote angiogenesis and tumor proliferation [18]. Moreover, myeloid-derived suppressor cells (MDSCs) decrease the activity of antitumor cytotoxic lymphocytes, accelerating tumor growth [19]. In 2014, Mabuchi et al. showed that uterine cervical cancer patients with elevated WBC counts were younger and had larger tumors and advanced stages. Moreover, TRL was associated with poor OS and resistance to radiotherapy, and TRL (+) patients showed significantly increased expression of granulocyte colony-stimulating factor (G-CSF) [7]. G-CSF-derived tumor increases MDSC in the bone marrow, blood stream, and spleen; this promotes a favorable microenvironment for tumor growth. Although the connection between MDSC and tumor growth has not yet been identified, other studies revealed that elevated levels of MDSC are associated with tumor proliferation. Thus, employing anti-MDSC agents could improve tumor response to radiation. Additionally, given that the spleen supplies MDSCs (hence the designation “splenic reservoir”), suppression of MDSC levels via splenectomy impairs G-CSF-derived tumor progression [7, 20, 21]. This study confirms the clinical significance of TRL and NLR in a large uterine cervical cancer patient population. Unfavorable prognostic features such as younger age, advanced FIGO stage, larger tumor size, and more LN metastases were more frequently accompanied by TRL. TRL (+)
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
However, the tumorigenic mechanism has not been clearly elucidated, although it appears likely that
11
patients also showed poor responses to radiotherapy and lower CR rates despite receiving higher doses, resulting in decreased OS; TRL remained a strong prognostic and predictive factor after performing PSM. Additionally, bone marrow hypermetabolism on FDG-PET occurred more frequently in TRL (+) patients, supporting the hypothesis that radioresistance is associated with increased levels of G-CSF-induced MDSC. Of note, a high NLR strongly correlated with poorer OS. In a previous study, patients with high NLRs showed low responses to local RT and granulocyte-macrophage colony-stimulating factor.
results are consistent with the notion that a high NLR is associated with the poor systemic disease control by suppressing the cytolytic activity of activated effector T cells as well as peritumoral infiltration of immunosuppressive macrophages [19]. Our findings strongly suggest that novel treatment strategies are required for certain patients. Namely, patients with TRL showing leukocytosis or neutrophilia at baseline, as well as those with bone marrow hyperactivity on PET, an enlarged spleen, or elevated WBC counts should be closely monitored during treatment, including their response to radiotherapy. Intensified radiotherapy (including simultaneous integrated boost or dose escalation) may improve local control for these patients. Moreover, systemic MDSC suppression treatment may be beneficial. A drawback of this study is that it was performed retrospectively over a long period of time. Furthermore, the selection of treatment modalities and regimens were heterogeneous throughout this period. Therefore, our findings should be interpreted with some caution. Conclusions Uterine cervical cancer with TRL is aggressive in nature and shows a poor response to radiation therapy. Our large population study validated the notion that both TRL and high NLR are associated with poorer outcomes. While further studies are required to investigate the biological mechanisms of TRL-promoted aggressiveness, our evidence clearly indicates that alternative
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Additionally, such patients did not benefit from the abscopal effect [22]. Along the same lines, our
12
diagnostic and treatment strategies are warranted for such patients.
ACKNOWLEDGEMENTS: None FUNDING: There was no research funding for this study. DISCLOSURE: The authors have declared no conflicts of interest relevant to this manuscript.
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
13
References 1.
Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer J Clin 2015; 65: 5-29.
2.
Oh CM, Won YJ, Jung KW et al. Cancer Statistics in Korea: Incidence, Mortality, Survival,
and Prevalence in 2013. Cancer Res Treat 2016; 48: 436-450. 3.
Annual report of cancer statistics in Korea in 2013. In. Korea Central Cancer Registry 2015.
4.
Kapp DS, Fischer D, Gutierrez E et al. Pretreatment prognostic factors in carcinoma of the
uterine cervix: a multivariable analysis of the effect of age, stage, histology and blood counts on survival. Int J Radiat Oncol Biol Phys 1983; 9: 445-455. 5.
Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature 2008; 454:
436-444. 6.
Granger JM, Kontoyiannis DP. Etiology and outcome of extreme leukocytosis in 758
nonhematologic cancer patients: a retrospective, single-institution study. Cancer 2009; 115: 39193923. 7.
Mabuchi S, Matsumoto Y, Kawano M et al. Uterine cervical cancer displaying tumor-
related leukocytosis: a distinct clinical entity with radioresistant feature. J Natl Cancer Inst 2014; 106. 8.
Lee YY, Choi CH, Kim HJ et al. Pretreatment neutrophil:lymphocyte ratio as a prognostic
factor in cervical carcinoma. Anticancer Res 2012; 32: 1555-1561. 9.
Teramukai S, Kitano T, Kishida Y et al. Pretreatment neutrophil count as an independent
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
http://ncc.re.kr/main.ncc?uri=english/sub04_Statistics
14
prognostic factor in advanced non-small-cell lung cancer: an analysis of Japan Multinational Trial Organisation LC00-03. Eur J Cancer 2009; 45: 1950-1958. 10.
Choi J, Yoon HI, Lee J et al. Optimal Extent of Prophylactic Irradiation of Paraaortic Lymph
Nodes in Patients with Uterine Cervical Cancer. PLoS One 2015; 10: e0145158. 11.
Kim KH, Kim S, Kim GE et al. De-escalation of the cumulative central radiation dose
outcome in patients with uterine cervical cancer. Gynecol Oncol 2015; 139: 439-446. 12.
Kim YB, Cho JH, Keum KC et al. Concurrent chemoradiotherapy followed by adjuvant
chemotherapy in uterine cervical cancer patients with high-risk factors. Gynecol Oncol 2007; 104: 58-63. 13.
Park TK, Choi DH, Kim SN et al. Role of induction chemotherapy in invasive cervical
cancer. Gynecol Oncol 1991; 41: 107-112. 14.
Mabuchi S, Matsumoto Y, Hamasaki T et al. Elevated white blood cell count at the time of
recurrence diagnosis is an indicator of short survival in patients with recurrent cervical cancer. Int J Gynecol Cancer 2012; 22: 1545-1551. 15.
Kim YB, Lee IJ, Kim SY et al. Tumor heterogeneity of FIGO stage III carcinoma of the
uterine cervix. Int J Radiat Oncol Biol Phys 2009; 75: 1323-1328. 16.
Mabuchi S, Matsumoto Y, Isohashi F et al. Pretreatment leukocytosis is an indicator of
poor prognosis in patients with cervical cancer. Gynecol Oncol 2011; 122: 25-32. 17.
Schetter AJ, Heegaard NH, Harris CC. Inflammation and cancer: interweaving microRNA,
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
according to the tumor response can reduce rectal toxicity without compromising the treatment
15
free radical, cytokine and p53 pathways. Carcinogenesis 2010; 31: 37-49. 18.
Gupta RA, Tejada LV, Tong BJ et al. Cyclooxygenase-1 is overexpressed and promotes
angiogenic growth factor production in ovarian cancer. Cancer Res 2003; 63: 906-911. 19.
Templeton AJ, McNamara MG, Seruga B et al. Prognostic role of neutrophil-to-
lymphocyte ratio in solid tumors: a systematic review and meta-analysis. J Natl Cancer Inst 2014;
20.
Abrams SI, Waight JD. Identification of a G-CSF-Granulocytic MDSC axis that promotes
tumor progression. Oncoimmunology 2012; 1: 550-551. 21.
Galdiero MR, Garlanda C, Jaillon S et al. Tumor associated macrophages and neutrophils
in tumor progression. J Cell Physiol 2013; 228: 1404-1412. 22.
Golden EB, Chhabra A, Chachoua A et al. Local radiotherapy and granulocyte-
macrophage colony-stimulating factor to generate abscopal responses in patients with metastatic solid tumours: a proof-of-principle trial. Lancet Oncol 2015; 16: 795-803.
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
106: dju124.
16
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
17
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Table 1. Patient and treatment characteristics All patients Characteristic
No.
TRL (-) group (%)
No.
TRL (+) group (%)
No.
n=2456
n=2058
n=398
56 (21-87)
56 (21-87)
53 (25-86)
(%)
P
Age Median (range)
<0.001
Clinical stage 494
20.1
465
22.6
29
7.3
IIA-IIB
1530
62.3
1288
62.6
242
60.8
IIIA-IIIB
394
16
280
13.6
114
28.6
IVA
38
1.5
25
1.2
13
3.3
Histologic subtype SCCa
2263
92.1
1988
92.2
365
91.7
non-SCCa
193
7.9
160
7.8
33
8.3
Tumor size (mm) Median (range)
40 (2-111)
40 (2-105)
50 (6-111)
<0.001
0.726
<0.001
<40
1046
42.6
959
46.6
87
21.9
40-50
570
23.2
490
23.8
80
20.1
≥51
840
34.2
609
29.6
231
58
Yes
516
21
394
19.1
122
30.7
No
1940
79
1664
80.9
276
69.3
≤10
207
8.4
145
7
62
15.6
>10
2249
91.6
1913
93
336
84.4
LN metastasis <0.001
Pre Tx Hb (mg/dl) <0.001
NLR Median (range)
2.17 (0.2-48)
2 (0.2-31.7)
4.3 (0.7-48)
<0.001
≤2.5 >2.5
1341
58.8
1260
66.1
81
21.5
940
41.2
645
33.9
295
78.5
Median (range)
4,088
3,762
8,024
(632-3,5512)
(632-3,5512)
(842-3,1574)
BM hypermetabolism Yes
59
32
17.6
27
62.8
No
166
150
82.4
16
37.2
ANC <0.001
<0.001
Treatment RT alone
1538
62.6
1322
64.2
216
54.3
Induction chemo +RT
333
13.6
266
12.9
67
16.8
CCRT
585
23.8
470
22.8
115
28.9
<0.001
Point A dose* Median
77.6
76.9
80.3
(range)
(51.7-122.2)
(51.7-122.2)
(59.1-101.6)
Duration of RT Median (range)
72 (36-181)
71 (36-171)
76 (45-181)
<0.001
<0.001
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
IA-IB
RT response CR
2339
95.2
1981
96.3
358
89.9
Non CR
117
4.8
77
3.7
40
10.1
<0.001
Abbreviation: SCCa; squamous cell carcinoma, LN; lymph node, Pre Tx Hb; pre-treatment hemoglobin, NLR: neutrophil/lymphocyte ratio, ANC; absolute neutrophil count, BM; bone marrow, RT; radiotherapy, CCRT; concurrent chemoradiotherapy CR; complete remission *EQD2, α/β=10
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Table 2. Patterns of failure according to tumor-related leukocytosis (TRL) and neutriophil/lymphocyte ratio (NLR) TRL (-) group
TRL (+) group
Low NLR
n = 2058
n = 398
n = 1341
No.
%
No.
%
P
No.
%
391
19
161
40.5
<0.001
256
19.1
Locoregional
234
11.4
112
28.1
<0.001
152
11.3
Distant
230
11.2
71
17.8
<0.001
149
11.1
Any recurrence
High NLR n = 940
No.
%
P
260
27.7
<0.001
164
17.4
<0.001
139
14.8
0.009
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
Table 3. Univariate and multivariate analysis for locoregional failure-free survival (LFFS) and overall survival (OS)
Variable
HR
LFFS
LFFS
OS
Univariate analysis
Multivariate analysis
Univariate analysis
95%CI
P
HR
95%CI
P
HR
95%CI
Age <56
1
≥56
0.688
1 0.055-0.853
0.001
0.72
1 0.661-0.975
0.029
0.879
0.734-1.053
Clinical stage IA-IB
1
1
1
OS Multivariate analysis P
HR
95%CI
P
0.161 1
IIA-IIB
2.787
1.865-4.164
<0.001
2.331
1.524-3.567
<0.001
2.099
1.566-2.831
<0.001
1.773
1.285-2.448
<0.001
IIIA-IIIB
6.963
4.555-10.643
<0.001
4.166
2.58-6.726
<0.001
5.85
4.24-8.071
<0.001
3.533
2.442-5.110
<0.001
IVA
9.731
4.821-19.644
<0.001
6.46
3.0.71-13.59
<0.001
10.548
6.275-17.73
<0.001
6.604
3.785-11.523
<0.001
SCC
1
1.318-2.351
<0.001
1.225-1.865
<0.001
1.618-2.45
<0.001
0.843-1.537
0.398
1.039-2.316
0.032
Histology Non-SCC Tumor size (mm) <40 ≥40
1.568
1 1.124-2.189
0.008
1 1.791
1.782
1 1.264-2.513
0.001
1 1.308-2.452
<0.001
1.619
1.23-1.977
<0.001
1.265
1.504
1.133-1.966
1 1.258-2.084
<0.001
2.546
0.972-1.647
0.081
2.363
0.005
1 1.76 1
2.219-3.044
<0.001
1.512
1.961-2.846
<0.001
1.991
LN metastasis No
1
Yes
1.56
Pre Tx Hb (mg/dl) <10
1
1
>10
0.654
No
1
Yes
2.944
1
1 0.469-0.913
0.018
1.189
1
1 0.816-1.731
0.367
0.563
1 0.43-0.736
<0.001
1.138
Leukocytosis
NLR
1 2.35-3.689
<0.001
2.103
1 1.615-2.737
<0.001
2.308
1 1.885-2.826
<0.001
1.552
1
>2.5
1.741
<0.001 1.396-2.171
1 1.065
1 0.830-1.365
0.621
1.897
1.576-2.284
ANC <6,000
1
≥6,000
2.808
1 2.266-3.542
<0.001
1.264
1 0.780-2.047
0.342
2.217
1.806-2.722
Treatment RT alone
1
1
1
Downloaded from http://annonc.oxfordjournals.org/ at University of Cambridge on August 13, 2016
≤2.5
<0.001
<0.001
1 1.305
1.051-1.62
0.016
0.626-1.443
0.81
1 0.95 1
Induction CRT
1.739
1337-2.261
<000.1
1.280
0.963-1.701
0.089
1.538
1.214-1.949
<0.001
1.023
0.794-1.318
0.861
CCRT
0.745
0.561-0.989
0.042
0.611
0.448-0.832
0.002
1.016
0.817-1.264
0.887
0.762
0.6-0.969
0.027
Abbreviations: SCCa; squamous cell carcinoma, LN; lymph node, Pre Tx Hb; pre-treatment hemoglobin, NLR: neutrophil/lymphocyte ratio, ANC; absolute neutrophil count, Induction CRT; induction chemotherapy followed by radiotherapy, CCRT: concurrent chemoradiotherapy