Thymidine kinase 1 in serum for prognosis, an investigation on routine individual adapted treatment of non-small cell lung carcinoma.

Background: Thymidine kinase 1 (TK1) is a key enzyme involved in DNA synthesis. The aim is to assess the prognostic significance of serum TK1 protein concentration (STK1p) and its role in monitoring of individual customized therapy in non-small cell lung carcinoma (NSCLC) patients in routine clinical setting. Methods: A prospective study of 129 NSCLC patients was confirmed by imager, /pathology and treated by radical resection (RR) combined with individual-chemotherapy or individual-chemotherapy alone, in 2010 to 2017. The STK1p was measured using an ECL dot blot assay in sera of patients. Results: Comparisons between 2-cycle-post-individual-chemotherapy and pre-individual-chemotherapy showed that STK1p was significantly associated with treatment effect (p<0.001), and the STK1p low-group correlated significantly to the early/middle clinical stage, as well as the treatment with RR+ individual - chemotherapy (p < 0.05). A significantly poor overall survival (OS) was found in elevated-risk STK1p vs. low-risk STK1p values chemotherapy alone vs. RR combined with individual - chemotherapy (p=0.001). The multivariate analysis showed that STK1p (hazard risk=2.295, p=0.010), and RR combined with individual-chemotherapy (hazard risk=3.04, p=0.0001), were independently survival factors. Conclusions: STK1p correlates significantly to survival and is an independent multivariate prognostic factor in NSCLC patients. STK1p as a low-cost assay, is a useful tool to combine with imager for a rational approach to increase the efficacy in early detection of tumor in lung cancer screening and assessment of individual adjusted therapy in NSCLC patients. Trial registration: No trial registration since this is not a case-control-study, but based on routine clinical work.


Background
Lung cancer is the top cause of mortality of both men and women in the world. Often, the symptoms of lung cancer do not appear until the cancer has been advanced, thus making the early diagnosis difficult. The 5-year survival rate for all people with lung cancer is 19% [1]. According to the 2015 report of cancer registration in China, the average 5 years survival rate was only 15% [2]. Although the false-positive lowdose computed tomography (LDCT) results in overloading of existing lung cancer clinics and multidisciplinary teams, a two large randomized controlled trials showed that early detection with LDCT and early-stage treatment can reduce mortality [3]. The USA Cancer statistics in 2019 reported that the decline in cancer mortality over the past 2 decades was primarily the result of steady reductions in smoking and advances in early detection and treatment, which were reflected in the rapid declines for the 4 major cancers (lung, breast, prostate, and colorectum) [1]. It proposals that complementation of LDCT examination with blood-based biomarkers is a rational approach to increase the efficacy for detection and treatment of early lung cancer in screening programs [4,5,6].
A series of serum/plasma biomarkers for routine setting are available for lung diseases, such as carcinoembryonic antigen (CEA) and cytokeratin 19 fragment (CYFRA21) [7]. Combination of CEA + CA125, or CEA + CY211 were effective for identifying suitable markers for lung cancer in screening [8]. The minimally invasive method of PD-L1 expression on circulating tumor cells (CTCs) can be performed instead of biopsy [9]. Thymidine kinase 1 (TK1) converts deoxythymidine (dTdR) to deoxythymidine monophosphate (dTMP) and its expression is closely related to cell cycle since 1950s [10]. TK1 is a precision protein molecular target for assessment of tumor proliferating rate [11,12,13,14,15,16]. The level of serum TK activity [17,18,19,20,21] and serum TK1 protein (STK1p) concentration [18,19,20,21] in malignant tumor patients is proportional to tumor proliferation, almost undetectable in normal human serum has been reviewed. We summered previous studies that STK1p could be a reliable tool in precise medicine for risk assessment of pre-malignancy and small invisible-malignancy in health screening, and for monitoring therapeutic effect in 14 type of tumors and prognostic factors of breast, non-Hodgkin's lymphoma and colorectal cancers in routing clinical oncology [22].
A meta-analysis of lung diseases showed that the expression of STK1p increased significantly in the following manner: healthy control group (n=1,694) < benign group (n=268) < carcinoma group (n=2,107) (p<0.00001) [23]. Similar results were obtained when measured the serum TK-activity and presented an independent prognostic factor [24]. The serum thymidine kinase 1 was an early detection marker in lung cancer using ELISA assay [21]. The ROC curve analyses showed that the STK1p combined with CEA, CYFRA21-1 and NSE was significantly higher than that of each biomarker alone (p < 0.0001) in lung cancer [25].
Most standard treatment protocols for lung cancer include cytotoxic agents, which are potential modulators of STK1 [24]. In this prospective study, we investigated if STK1p could be used to assist individual customized chemotherapy treatment efficacy in NSCLC patients (n=129) with early/middle and advantage stage.

Patients
A prospective cooperative study of total 129 patients with non-small-cell-lung- This is not a case-control-study, but based on routine clinical work, where serum samples were collected from time to time, pre-chemotherapy and after 2 cyclechemotherapy (table 1 and table 2), depending on patient's medical situation, for more details, see section "Treatment" below. According to previous publications it was found that some of patients with advanced stage (IIIB-IV) had lower STK1 levels compared to patients of early/ middle stage. Therefore, the STK1 values prechemotherapy were not used for statistical assessment of prognosis of overall survival (OS) and COX analysis.

STK1p in relation to cycles of chemotherapy
The STK1p value 3-month-post-treatment was a best time to evaluate the prognosis of recurrence risk in breast cancer patients, which has been recommended [26]. In a previous study on NSCLC patients it was found that the TK-activity in serum (STKactivity) after 1-2 cycles of chemotherapy correlated to survival [24]. In order to find a suitable cycle chemotherapy when STK1p expression was significantly correlated to survival rate. We performed an initially testing in advantage NSCLC patients before and 2, 4, 6 and 8 cycles after chemotherapy in the oncology department of the First Affiliated Hospital of Suzhou University. We used CYFRA21-1 as a control.
The results are summarized in table 2. We found that the STK1p values correlated significantly to survival after 2-cycle individual-chemotherapy (p< 0.05), but not after 4, 6 or 8 individual-chemotherapy cycles. Therefore, we used the STK1p values after 2 chemotherapy-cycles in our clinical study presented here.

Treatment
According ECT (electro-chemotherapy) was used for brain metastases.

Follow up
The primary endpoint was progression-free survival (PFS) and the secondary endpoint was overall survival (OS), during the follow-up period of 10-58 months. The criteria included was STK1p levels, age, sex, clinical stage, pathological type (AC and SCC), and RECIST. The criteria excluded was patient's ID, phone number, home address, data when visiting/left the hospital, routine blood and urine test results. The patients or their families were contacted by phone every month to check if they remained alive. The survival assessment was performed only in 101 patients, due to limited access to 28 patients once they left the hospitals.

STK1p assay
STK1p concentrations were determined using a commercial kit (cat. no. 24/48T based on an enhanced chemiluminescent dot blot assay as described by the manufacturer Sino-Swed Tongkang Bio-Tech, Inc., Shenzhen, China, www.sstkbiotech.com). The collection of serum was performed between 7.30 am-10.00 am after the patients had fasted for 12-14 h. The samples were analyzed within 3 hrs.

STK1p levels before and after 2 cycle-chemotherapy
The patients were divided into four groups: total number of patients of all clinical stages (n=129), follow-up patients of stage I-IIIB (n=34), follow-up patients of advance stage (IIIB+IV, n=71) and follow-up patients of all stages (n=101) ( Table 3).
The pre-treatment mean STK1p value of malignant patient's is normally above 2.0 pM and the STK1p of healthy people is less than 0.5 pM (Chen et al., 2011. In this study, however, the treatment situation was complex. Before start of the type of individual-chemotherapy we focused on in this study, the all patients with clinical early/middle stage and 30.3% patients with clinical advantage stage had already received surgery and in some cases also traditional Chinese medicine treatment, reducing the mean STK1p value to about 1.0 pM and continue to significantly decrease during the individual-chemotherapy to about 0.7 pM (table 3), but still significantly higher than the normal control of 0.38 pM (table 3).
The STK1p-individual values decreased by about 30%, except for the stage I-IIIA patient group because STK1p was already decline due to radical surgery. Individual examination shows that in about 40% of the patients the STK1p value decreased by 50% or more at the end of the chemotherapy (table 4), indicating an efficient treatment, including patients with advanced stage. It should be noted that the mean STK1p value is higher in the advance stage (IIIB-IV, 1.15pM) group compared to the early stage (I-IIIA, 0.79 pM) group, although not significant different (p=0.157) ( Table 5).

The number of STK1p low-and high-risk patients in relation to clinical characteristics 2-cycle-post individual-chemotherapy
To be able to identify low and high-risk groups of patients a suitable threshold level individual-chemotherapy alone (p < 0.05), but did not significantly correlated to pathology type, age, sex and RECIST, and the treatment regimen of advanced stage (IIIB+IV) only (p > 0.05). The STK1p low risk and elevated risk groups did not significantly correlated to smoking history (p = 0.207, data not show)

The overall survival (OS) rate
The total mortality was 39.6%. The patient at early/middle and at advantage stage was 11.5% and 49.3%, respectively.
The overall survival (OS) rate in relation to stage between low risk and high risk of STK1p group, I-IIIA and IIIB-IV, pathological type between AC and SCC and individual therapy program with RR+ individual-chemotherapy compared with individual-chemotherapy alone are shown in figure 1. There was a significantly higher OS in patients of low risk compared to patients of high risk (p=0.015) (Fig1 A), of early/middle stages compare to advanced stage (p=0.011) (Fig. 1 B), of AC patient compare to SCC patients (p=0.028) (
The multivariate analysis showed that the treatment (RR+ individual-chemotherapy vs. individual-chemotherapy alone, p = 0.001) and the STK1p expression (p = 0.01) are independent prognostic factors for OS, while stage and pathology type and sex were not (p> 0.05, table 6).

Cases analysis of early stage dead patients
Although the patients of early stages (I-IIIA) showed a good prognosis with high OS rate (table 1,

Discussion
In In the present study, we chose to use STK1p values after 2 cycle-individualchemotherapy, since it correlated closely to the endpoint of the treatment, supported by a similar study of STK activity in NSCLC patients [25]. The mean STK1p values decreased by about 30%, except for the stage I-IIIA patient group. In addition, individual examination shows that in about 40% of the patients the STK1p value decreased by 50% or more after the 2-cycle-chemotherapy (table 4), indicating rather efficient individual treatment, including patients with advanced stage. Therefore, we investigated The end of the 2-cycle individual-chemotherapy is a useful time-point when using STK1p for monitor treatment with these agents in NSCLC patients. However, the STK1p values in about 60% of the patients at the end of the 2-cycle individualchemotherapy were still significantly higher (~0.70 pM) compare to normal health persons (0.38 pM), it means that some residual tumors are growing and might result that STK1p did not return to normal levels as described in 3.6, cases analysis of early stage dead patients. The normal STK1p value of 0.3-0.5 pM was confirmed in a 160,086-participants-health screening [27].
In this study, we divided the STK1p 2-cycle-individual-chemotherapy patients into a low risk (n=58) and an elevated risk (n=43) groups and compared them with parameters of clinical characteristics. The only significantly difference was found that the STK1p values decreased significantly in the early/middle stage patients, but not in the patients with advance stages. The STK1p values also decreased significantly after surgery + individual-chemotherapy in the early/middle stage patients compared to individual-chemotherapy only, but not in the advanced stage patients, indicating that the STK1p low-risk value was significant correlated to the efficient combined treatment in the early/middle stage patient, but not in the advanced stage patients.
When considered elevated STK1p value, advantage stage, pathologic type of SCC and individual-chemotherapy alone correlated to poor OS in the Kaplan Meyer curves STK1p (≦0.6 vs. ＞0.6 pM, p=0.01) were independent prognostic factors, while stage, pathology type and sex were not (p> 0.05). The present study is consistent with a previous retrospective TK-activity [24] and immunohistochemical studies on patients with lung cancer [28], the high TK1 expression was independent prognostic factor of poor survival.
Generally speaking, early/middle stage patients showed a better survival. However, there was three patients who died within 14-20 months after end of the chemotherapy.
The STK1p values of the post-individual-chemotherapy did not returned to normal values (<0.6 pM), and were characterized as RECIST SD. The CT examination showed that these patients were already in a metastatic stage, and it was may be too late to start chemotherapy.
Although trying to implement a better individual customized therapy program, the Kaplan Meyer curves showed that the OS rate of the advantage stage (IIIB+IV) patients displayed slightly improvement, but not significant different between patients with low STK1p values compared to patients with high STK1p values (Fig 1 H, p=0.13). The mortality was still high (49.3%). In a previous study, it was reported that STK1p was low in 63.5 % of lung cancer patients with advantage stage with increased tumor size [29]. The high degree of necrosis is normally found in patients with stage III/IV, the decline STK1p was not corresponding to tumor size of patients at advantage stage [30].
A study on STK-activity (TKa) and circulating tumor M2 pyruvate kinase (TuM2-PK) concentration in renal cell carcinoma (RCC) demonstrated high degree necrosis in RCCs with stage III/IV. The presence of extensive tumour necrosis (>50%) was statistically correlated to high TuM2-PK concentration and low STKa [31]. A likely explanation is that advantage stage patients already had low STK1p when started the chemotherapy due to extensive necrosis with reduced growth rate [31]. Thus, it should be noted that STK1p is may not a suitable prognostic marker when there is a risk of extensive necrosis in advantage stage of lung patients. Early detection and early treatment would lead to rapid declines in mortality [21,22].
In this study, it was obvious that surgery in combination with individualchemotherapy was a more efficient than individual-chemotherapy alone. Furthermore, in the COX analysis it was found that the surgery + individual-chemotherapy was an independent prognostic factor for survival. More recently, due to the immunotherapy revolution, specifically the development of immune checkpoint inhibitors (ICIs), such as PD-L1 expression, can improve the treatment of advantage lung cancer patients [32].
A meta-analysis revealed that immunotherapy in combination with chemotherapy is an effective option as a first-line treatment for lung cancer (n = 4,887) [33]. The NSCLC patients administered immune-and chemotherapy exhibited better PFS and OS than patients treated with chemotherapy alone (P<0.001). Moreover, as the expression of PD-L1 increased, the PFS and OS benefits were more significantly [34]. The CTCs from peripheral blood can provide a minimally invasive method to monitor PD-L1 expression on tumor cells over time [9], but the protocol is complicated and timeconsuming. As an alternative to use biopsy or CTCs to measure PD-Li expression, we suggested to use STK1p assay, a non-invasive, simple, and low cost assay, for monitoring the effect of the treatment. Of NSCLC, two most common subtypes are adenocarcinoma (AC, ∼70%) and squamous cell lung cancer (SCC, ∼30%) [35]. In this study, the OS was significant better in AC patients compared to SCC patients indicating that the SCC type of patients contain high-risk individuals. This is in coincide with a previous study reported that STK1 level was significantly higher in SCC as compared to AC patients [36]. In addition, in a recent study, extensive plasma lipidomics profiling of patients with SCC was performed and a panel of 2 lipid biomarkers that were able to detect SCC during the early stages were identified. The results of that study revealed the potential of using these 2 lipid markers in combination with LDCT-based screening methods in order to identify SCC patients with high-risk [37]. It is possible that STK1p in combination with such 2 lipid markers could improve the assessment of high-risk in early SCC patients.
Cancer is a chronic individual-disease with abnormal growth. Every cancer arises from a pre-cancer, although not all pre-cancers lead to cancer [22,38]. If all pre-cancer were detected early, the risk for further development into malignancy would be reduced extensively. A series of papers proved that STK1p level increases earlier than the discovery of the visible cancer by imaging [22]. For example, a case report showed that a woman (aged 72 years) was diagnosed by imaging as having ground-glass opacity (pre-cancer) of lung. STK1p value was high of 3.9 pM while CEA, Cyfra19 and NSE were normal. The STK1p continued to rise to above 10-13 pM, while CT scanner and the three markers were still normal 4 months after discovery of right lung cancer. A radical operation was performed and confirmed AC (moderate differentiation) by pathology. The STK1p value declined to 1.3 pM 1 month after surgery, no occurrence of metastasis or recurrence was detected by CT scanner 5 months after follow-up [27].
Precision molecular medicine with many biomarkers are either available or under development for several lung diseases. However, few biomarkers are in early widespread clinical use so far. It is mandatory to show the performance of these biomarkers for the monitoring and prognosis in early lung tumors, such, pre-cancers (ground-glass opacity preaches) or malignant tumor (asymptomatic malignant lung nodules, invisible small tumor or early localized tumors), which are potential to use for monitoring and prognosis of personalized targeted therapies. TK1 is one of a precise molecule target for monitoring and prognosis, specifically, for early detection of growth of invisible small tumors. STK1p assay is also non-invasive and low cost [22]. We proposal that STK1p combined with imaging and/or another biomarkers, for example CTCs and TK1 immunohistochemical staining, might be a more reliable assessment for developing of pre-malignancy or diseases associated with the risk process of individual lung cancer. There would be no or very limited lung cancer development in the future if the pre-malignancy or early tumors are treated in time.
Conclusions: STK1p, a low-cost assay, correlates significantly to survival and is an independent multivariate prognostic factor of NSCLC patients. The combination of the STK1p with suitable imager is a rational approach to increase the efficacy for early detection of tumor in lung cancer screening. An individual customized therapy, such as RR combined with chemotherapy, would reduce the mortality in routine clinical setting of NSCLC.  Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.