Sage Journals: Discover world-class research

Abstract

BACKGROUND:

People with biomarkers above cut-off values normally have higher risk to develop pre-malignancies and malignancies.

OBJECTIVE:

Here we investigate if serological TK1 protein (STK1p), AFP, CEA and PSA below cut-off values predict development of pre-cancer.

METHODS:

The mean values and the concentration distribution of STK1p, AFP, CEA and PSA were determined in a cohort of 56,178 persons participating a health screening group, consist of people with non-tumor diseases, pre-malignancy and diseases associated with the risk process of malignancy. A health disease-free group ( $n=$ 428) was selected among the 56,178 participants and used as controls.

RESULTS:

The STK1p below cut-off value ( $\leqslant$ 2 pM) showed partly (51.6%) an almost normal concentration distribution and partly (43.9%) an extensive tail in the health screening group, which was not found in the disease-free group. Due to the extensive tail in the distribution, the mean value of STK1p increased significantly ( $p=$ 0.0001) from 0.38 $\pm$ 0.30 pM in the health disease-free group to 0.69 $\pm$ 0.55 pM in the group below the cut-off value. No significantly differences in the concentration distribution and the mean values among gender and ages were observed. On the other hand, there were no difference in the concentration distributions and the mean values of AFP, CEA and PSA between the health disease – free group and the group below cut-off values, as well as between gender and ages. Of interest, the elevated mean value of STK1p of the group below the cut-off value was correlated to pre-malignancy and diseases associated with the risk process of malignancy in liver and prostate. No such correlations were found with AFP, CEA and PSA.

CONCLUSION:

STK1p is a potential proliferating biomarker for early discover of persons in the risk to develop or already have pre-malignancies or diseases associated with the risk process of malignancy.

Keywords

Thymidine kinase 1 (TK1)serum thymidine kinase 1 (STK1p)AFP CEA PSA pre-malignancy

1. Introduction

Cancer is a leading cause of death worldwide. An estimated 40 million deaths occurred in 2015 due to non-communicable diseases (NCDs), accounting for 70% of the overall total of 56 million deaths. The majority deaths were caused by the two main NCDs, 17.7 million deaths (45%) of cardiovascular disease and 8.8 million deaths (22%) of cancer [1]. In China, lung, liver, gastric, oesophageal, colorectal, breast and pancreatic cancers are the leading causes of cancer deaths [2, 3]. However, since there is a great geographical span, an uneven economic development, and differences of natural environments influencing lifestyle and eating habits, the types of cancer diseases are rather different in different parts of China, even within local areas. For example, in a study on a cohort in 486,085 people of a routine health screening at the Health Centre, Fujun 180 Hospital, Quanzhou city, south-east part of China, the incident rate of cancer increased rapidly by about five times from 2009 to 2014. The most frequently cancers were lung and liver, and in female cervix cancer. Of the 56,286 persons included in this health screening, 89 persons were discovered with malignancies and of these patients 52 were followed-up during 10–40 months. The average mortality was 48.5%, while the mortalities of the liver and lung cancer patients were 94% [4].

World Health Organization (WHO) reported that during the next 20–25 years the number of new cancer patients world-wide will increase by about 70%, from today of 14 million to about 25 million, i.e. we lost the “war” against the cancer, if we just continue as now. This is not only a tragedy for individual persons, but also a dramatically increases in the expenditures for the health care system. If we can prevent this, we not only help people with the risk of developing cancer, but also reduce the burden in the medical health care system [5]. How do we follow the strategy of WHO “Prevention and early detection of cancers”? Cancer is a long-term chronic and abnormal proliferating disease. Multiple gene mutations associated with cell growth regulation in certain enzymes and proteins leads to uncontrolled proliferation and thus development of malignancies, which may be take 10–30 years. Using tumor proliferating markers are may be the key point. Non-invasive serological methods for early detection of tumor may be give information of pre-malignant process and thus increases the possibility to cure patients. However, there are few tumor markers for early tumor screening in routine health centre today. Example of such markers is CEA (carcinoembryonic antigen) and AFP (alpha-fetaprotein) in commonly used for tumor screening. Since CEA and AFP markers are not directly assessing the rate of tumor proliferation [6], the value of early detection of cancers is still doubt.

Thymidine kinase 1 (TK1) was used as a precision protein molecular target for assessment of cell proliferating rate since 1950s. The TK1 level in tumor cells [7, 8, 9, 10, 11, 12], in tissues [11, 12, 13, 14, 15, 16, 17] or in serum [4, 11, 12, 18, 19, 20, 21, 22, 23] corresponds to cell proliferating rate and is expected to be a favour marker for assessment of tumor proliferating rate. Based on analysis of high number serum samples, serum TK1 concentration (STK1p) combined with early risk symptoms of tumor process, such as pre-cancer or tumor-related diseases, could be used as a strategy of early detection of cancers and thus for risk warning of development of malignancies later in the life [4, 11, 12, 20, 21, 22, 23]. According to a report of Wang et al. [4], elevated STK1p was more sensitive than CEA and AFP for patients who had malignant tumors and was also a reliable prognostic biomarker for death, while CEA and AFP were not.

To further understand how to use STK1p, CEA or AFP markers for early risk screening of pre-malignancy or diseases related cancer in the routing health centre, we determined the concentration distribution of STP1p, AFP and CEA in relation to ages in 56,178 persons participated a health screening. These persons had no symptoms of malignancies. The study was done at a health centre in south-east of China (QuanZhou, Fujian Province), an area with higher frequency of liver, lung and cervix cancers, during 2011–2014. This type of study might provide guidance for pre-cancer risk assessment, and are expected to lead to an early tumor detection in routine health screening.

Table 1
The mean and median values of STK1, AFP and CEA with progressive ages

A. STK1p (pM)	1. Health disease-free				2. Health screening (STK1p $\leqslant$ 2 pM)				3. Health screening (STK1p $>$ 2 pM)				1 vis. 2
Age	Meam $\pm$ SD	95%CI	Min $\sim$ Max	Median	Meam $\pm$ SD	95%CI	Min $\sim$ Max	Median	Meam $\pm$ SD	95%CI	Min $\sim$ Max	Median	Mean value $p$
21–40 ys.	0.22 $\pm$ 0.11	0.17 $\sim$ 0.26	0.10 $\sim$ 0.54	0.18	0.71 $\pm$ 0.57	0.70 $\sim$ 0.72	0.01 $\sim$ 2.00	0.52	3.79 $\pm$ 3.66	3.55 $\sim$ 4.02	2.01 $\sim$ 40.00	2.895	0.00
	( $n=$ 25)				( $n=$ 20535)				( $n=$ 952)
41–60 ys.	0.40 $\pm$ 0.32	0.37 $\sim$ 0.43	0.05 $\sim$ 1.82	0.33	0.67 $\pm$ 0.53	0.67 $\sim$ 0.67	0.01 $\sim$ 2.00	0.51	3.66 $\pm$ 2.40	3.60 $\sim$ 4.00	2.01 $\sim$ 40.00	2.97	0.00
	( $n=$ 363)				( $n=$ 28090)				( $n=$ 1046)
$>$ 60 ys.	0.32 $\pm$ 0.19	0.26 $\sim$ 0.38	0.11 $\sim$ 0.98	0.25	0.67 $\pm$ 0.52	0.66 $\sim$ 0.68	0.01 $\sim$ 2.00	0.52	3.81 $\pm$ 3.61	3.33 $\sim$ 4.30	2.01 $\sim$ 40.00	2.82	0.00
	( $n=$ 40)				( $n=$ 5331)				( $n=$ 213)
Summary	0.38 $\pm$ 0.30	0.35 $\sim$ 0.41	0.05 $\sim$ 1.82	0.31	0.69 $\pm$ 0.55	0.68 $\sim$ 0.69	0.01 $\sim$ 2.00	0.52	3.79 $\pm$ 3.48	3.65 $\sim$ 3.94	2.01 $\sim$ 40.00	2.91	0.00
	( $n=$ 428)				( $n=$ 53956)				( $n=$ 2211)
B. AFP (ng/ml)	1. Health disease-free				2. Health screening (AFP $\leqslant$ 10 ng/ml)				3. Health screening (AFP $>$ 10 ng/ml)				1 vis. 2
Age	Meam $\pm$ SD	95%CI	Min $\sim$ Max	Median	Meam $\pm$ SD	95%CI	Min $\sim$ Max	Median	Meam $\pm$ SD	95%CI	Min $\sim$ Max	Median	Mean value $p$
21–40 ys.	2.84 $\pm$ 1.82	0.37 $\sim$ 2.07	0.22 $\sim$ 8.11	2.425	2.74 $\pm$ 1.42	2.72 $\sim$ 2.76	0.00 $\sim$ 9.98	2.54	50.22 $\pm$ 136.18	30.20 $\sim$ 70.24	10.04 $\sim$ 1200	14.31	0.73
	( $n=$ 24)				( $n=$ 18732)				( $n=$ 182)
41–60 ys.	3.99 $\pm$ 1.98	3.72 $\sim$ 4.26	0.70 $\sim$ 12.67	3.63	3.29 $\pm$ 1.59	3.27 $\sim$ 3.31	0.00 $\sim$ 10.00	3.06	35.65 $\pm$ 136.80	22.37 $\sim$ 52.68	10.01 $\sim$ 1200	12.38	0.00
	( $n=$ 211)				( $n=$ 25632)				( $n=$ 309)
$>$ 60 ys.	3.53 $\pm$ 1.79	2.53 $\sim$ 4.52	0.65 $\sim$ 7.25	3.72	3.28 $\pm$ 1.54	3.24 $\sim$ 3.33	0.05 $\sim$ 9.95	3.08	70.61 $\pm$ 221.98	$-$ 4.49 $\sim$ 145.72	10.36 $\sim$ 1200	12.76	0.531
	( $n=$ 15)				( $n=$ 4401)				( $n=$ 36)
Summary	3.85 $\pm$ 1.97	0.12 $\sim$ 4.10	0.22 $\sim$ 12.67	3.54	3.08 $\pm$ 1.55	3.06 $\sim$ 3.09	0.00 $\sim$ 10.00	2.87	43.07 $\pm$ 144.18	32.15 $\sim$ 56.02	10.01 $\sim$ 1200	12.77	0.00
	( $n=$ 250)				( $n=$ 48765)				( $n=$ 527)
C. CEA (ng/ml)	1. Health disease-free				2. Health screening (CEA $\leqslant$ 5 ng/ml)				3. Health screening (CEA $>$ 5 ng/ml)				1 vis. 2
Age	Meam $\pm$ SD	95%CI	Min $\sim$ Max	Median	Meam $\pm$ SD	95%CI	Min $\sim$ Max	Median	Meam $\pm$ SD	95%CI	Min $\sim$ Max	Median	Mean value $p$
21–40 ys.	1.48 $\pm$ 0.91	1.09 $\sim$ 1.86	0.10 $\sim$ 4.02	1.43	1.65 $\pm$ 0.97	1.64 $\sim$ 1.67	0.00 $\sim$ 5.00	1.49	6.58 $\pm$ 2.31	6.30 $\sim$ 6.85	5.01 $\sim$ 28.97	5.96	0.391
	( $n=$ 24)				( $n=$ 18149)				( $n=$ 268)
41–60 ys.	2.11 $\pm$ 1.24	1.94 $\sim$ 2.29	0.22 $\sim$ 10.04	1.915	1.92 $\pm$ 1.03	1.91 $\sim$ 1.93	0.02 $\sim$ 5.00	1.76	8.23 $\pm$ 16.93	7.21 $\sim$ 9.26	5.01 $\sim$ 437.96	6.31	0.01
	( $n=$ 198)				( $n=$ 24573)				( $n=$ 1050)
$>$ 60 ys.	2.77 $\pm$ 1.95	1.64 $\sim$ 3.89	0.21 $\sim$ 7.88	2.235	2.31 $\pm$ 1.10	2.27 $\sim$ 2.37	0.06 $\sim$ 5.00	2.19	16.58 $\pm$ 86.20	7.48 $\sim$ 25.69	5.01 $\sim$ 1000	6.28	0.12
	( $n=$ 14)				( $n=$ 4051)				( $n=$ 347)
Summary	2.09 $\pm$ 1.28	1.92 $\sim$ 2.25	0.10 $\sim$ 10.04	1.86	1.85 $\pm$ 1.03	1.84 $\sim$ 1.86	0.00 $\sim$ 5.00	1.68	10.89 $\pm$ 47.94	8.60 $\sim$ 13.18	5.01 $\sim$ 1000	6.24	0.00
	( $n=$ 236)				( $n=$ 46773)				( $n=$ 1686)

2. Participants and methods

2.1 Participants

In 2016, we reported a retrospective study on 56,286 people participated in a health screening during 2009–2014. Eighty-nine persons were found to have malignancies of different types [4]. Here we further analysis data from this study by focus on those persons ( $n=$ 56,178; men 38,816, women 17,362) that have no symptom of malignancies but diseases linked to development of tumors (pre-cancer) to see how well biomarkers could predict pre-cancer. We determined the mean values and the distribution of the concentrations of STK1p, AFP, CEA and PSA, the mean age of the total population was 45.7 $\pm$ 10.9. Of those 56,178 persons, 428 were found to have no symptoms of pre-malignancies/malignancies or other diseases and used as health controls. The health screening study was performed at the 180 Hospital Health Centre, Quanzhou, Fujian, China, during 2009–2014. The study was approved by the Committee of Research Ethics at 180 hospital, Quanzhou, Fujian, China. All the persons gave informed consent to participate in this study.

All persons were covered by medical insurance and represented mainly people living and working in this area. A routine health screening system was used, including ultrasound, X-ray, electrocardiography, hepatitis-B virus infection test, blood/urine routine testing, physical examination, etc. In addition, 56,178 were tested for STK1p, 49,441 for AFP, 48,771 for CEA and 13,897 for PSA. In the health control group, 428 were tested for STK1p, 250 persons for AFP, 236 for CEA and 30 for PSA.

The participants were divided into three age groups: 21–40, 41–60 and $>$ 60 years (Table 1).

2.2 Methods

The assessment of pre-malignancy and diseases link to the risk of developing malignancies has not been standardized. The concept of pre-cancers are the earliest morphologically discernible lesions that present the development of invasive cancer. In fact, the term pre-cancer is often confused with many unrelated terms. Currently, the term “intraepithelial neoplasia” is wide usage by pathologists. There are hundreds of different pre-cancers of human. A more extensive listing of pre-cancers could be found in the appendix published by Berman and Moore [24]. In this study, we attempted to tailor multiple available options for assessment of pre-malignancy and diseases link to the risk of development of malignancies described by Berman and Moore in 2010.

2.2.1 Discovery of tumor-related symptom

In order to screen for pre-cancer or diseases link to developing malignancy or suspected malignancy, an advanced Doppler Ultrasound Diagnostic Apparatus (Voluson E8, American General Corporation) was used. For example: Liver and gallbladder diseases (liver fibrosis, liver cirrhosis, or suspicion of liver cancer); uterine diseases (tumors, endometriosis, uterine abnormalities, ovarian mass, pelvic inflammatory mass or abscess, etc.); breast diseases (the size, shape and edge of the lesion, it is an important and convenient method for early diagnosis of breast hyperplasia); prostate diseases (shape, size and position, the diagnosis of prostatic hyperplasia, calculus, calcification disease and prostatic abscess, cyst or chronic prostatitis, suspicion of prostate, cancer. It is also suitable for benign prostatic hyperplasia patients); thyroid diseases (size and volume, to assist the identification of pure, nodular, thyroiditis, hyperthyroidism, thyroid tumor such as adenoma, cyst, thyroid, parathyroid hyperplasia, cyst and parathyroid carcinoma). Additionally, it was also used for preliminary screening of heart diseases, such as internal structure, heart beat and blood flow, and other types of heart disease or cardiomyopathy.

The X-ray was combined with CT/SCT for further detection who had suspicion lung cancer.

2.2.2 Serological assays

All serum samples were collected in the morning on fasting condition. Blood collected without anticoagulant was centrifuged at 800 $\times$ g for 10 min, 2–3 hours after collection and transferred into tubes and kept at $-$ 20 ${}^{\circ}$ C before analysis. Persons expressing values of less than the cut-off values were denoted as “low-risk group”, regarded as a low risk of developing malignancies, while persons expressing values above the cut-off values were denoted as “high-risk group”, likely representing individuals with increasing risk of malignancy progression.

2.2.2.1. Serum TK1 assay

The concentration of serum TK1 (STK1p) was measured by using a commercial kit based on an enhanced chemiluminescent (ECL) dot blot assay as described by the manufacturer (SSTK Ltd., Shenzhen, China). Samples comprising 3 $\mu$ L of serum were directly applied to nitrocellulose membrane in duplicates. The serum samples were probed with anti-TK1 chicken IgY antibody raised against a peptide (residue 195–225, GQPAG PDNKE NCPVP GKPGE AVAAR KLFAPQ). TK1 peptide was dotted at different concentrations (2.2, 6.6 and 20 pM) as a quality control standard. The intensities of the spots on the membrane were determined by a CIS-l Imaging System (SSTK Ltd., Shenzhen, China). From the intensities of the TK1 quality control standard of known concentrations, the STK1p was calculated and expressed as pM. For detail description of the STK1p assay, see reference [21]. In this study, 2.0 pM of STK1p was used as a risk threshold value.

2.2.2.2. AFP and CEA assay

The levels of AFP and CEA were performed on an electrochemiluminescence automatic immunoassay (Roche Diagnostics e601GmbH, Mannheim, Germany). The reference cut-off values were 10.0 ng/mL and 5.0 ng/mL, respectively. The electrochemiluminescence (ECL) automatic immunoassay analyser is based on paramagnetic particles as a solid phase, biotin-streptavidin-detection system and two-dimensional bar code technology. It is a highly sensitive light detection system that provides excellent low-end sensitivity and broad dynamic measuring ranges.

Figure 1.

Concentration distribution of STK1p, AFP and CEA of health disease-free group and of people below cut-off values (low-risk group) in relation to progressive age.

Figure 1.

continued.

2.2.2.3. PSA assays

PSA was determined by an electrochemiluminescence immunoassay. The electrochemiluminescence automatic immunoassay analyzer is based on paramagnetic particles as a solid phase, biotin-streptavidin-detection system and two-dimensional bar code technology. It is a highly sensitive light detection system that provides excellent low-end sensitivity and broad dynamic measuring ranges. Assays of PSA used Beckman DX1800, Germany. The reference cut-off value was 4.0 ng/ml. The value was considered positive or negative for the marker when the level was above or below the cut-off value.

2.2.2.4. ALT, AST and GGT assays

Three enzymes in serum, alanine aminotransferase (ALT), aspartate aminotransferase (AST), $\gamma$ -glutamyl transferase (GGT) were detected by using a Toshiba TBA120FR automatic biochemical analyser. The kits were provided from the manufacturer Zhejiang Century Cornell Medical Technology Co. China. The normal reference range: ALT $\leqslant$ 40U/L, AST $\leqslant$ 40U/L, GGT $\leqslant$ 50 U/L. More than 20% higher as compared to the normal reference values is considered as liver dysfunction.

2.3 Statistical analysis

The mean values of STK1p, AFP and CEA levels were calculated by a mean $\pm$ standard deviation program (Microsoft Excel). For comparison of STK1p concentration levels among the different groups of persons investigated, T-test and Chi-square tests were used (IBM SPSS Statistics 19). P-values $\leqslant$ 0.05 were regarded as statistically significant.

3. Results

The persons participated in the health screening were divided into three age groups: 21–40, 41–60 and $>$ 61 years. Each age groups were further divided into three groups: 1. persons with no symptoms of pre- and malignancy or any other diseases linked to development of malignancy (health disease-free group); 2. persons below the cut-off value of STK1p, AFP, CEA and PSA (low-risk group); 3. persons above the cut-off value of STK1p, AFP or CEA (elevated-risk group). Since there was no significant difference in the STK1p, AFP and CEA values between men and women, the individual values of them were put together. The mean and median values of STK1p, CEA and AFP in the different groups are summary in Table 1.

Figure 2.

Correlations between STK1p and AFP, STK1p and CEA and between CEA and AFP in serum of people below cut-off values (low-risk group).

3.1 STK1p, AFP and CEA values in relation to ages

There was a significantly increase in the STK1p, AFP and CEA values by age in the health disease-free group, as well as in the low-risk AFP and CEA group (Table 1). No significant difference was found in the STK1p, AFP and CEA values in the elevated-risk group by age (Table 1).

3.2 Comparison of health disease-free and low-risk groups

We further analysis the results by compare the STK1p, AFP and CEA values of the low-risk group (below cut-off) with the health disease-free group to find out if those biomarkers could indicate risk for development of pre-cancer and in the end malignancy.

3.2.1 STK1p

There were significantly higher STK1p values in the low-risk group compared to the health disease-free group (Table 1A) in all age groups (21–40, 41–60, $>$ 60 years). Putting all age groups together the mean value of STK1p in the low-risk group was 0.69 $\pm$ 0.55 pM compared to 0.38 $\pm$ 0.30 pM in the health disease-free group ( $p<$ 0.0001) (Table 1A).

3.2.2 AFP

In opposite to STK1p, AFP values in the low-risk group ( $<$ 10.0 ng/ml) were lower in all age groups compared to the health disease-free group, although not significant (Table 1B). The mean value of all age groups together in low-risk group was 3.08 $\pm$ 1.55 ng/ml compared to 3.85 $\pm$ 1.97 ng/ml in the health disease-free group, showing an inverse significant difference (Table 1B).

Table 2
STK1p, AFP and CEA values in the low-risk group with different types of liver pre-malignancy diseases

Type of liver	Health	*Moderate/severe	*HBV infection	*Gallbladder	*Obesity	*Abnormal	*Cyst	*Echo	**Cirrhosis
diseases	disease-free	fatty liver	high risk	polyps		function		thickening
STK1p (pM)	0.38 $\pm$ 0.30	0.60 $\pm$ 0.51	0.59 $\pm$ 0.51	0.64 $\pm$ 0.54	0.58 $\pm$ 0.49	0.54 $\pm$ 0.48	0.63 $\pm$ 0.56	0.69 $\pm$ 0.56	0.96 $\pm$ 0.53
AFP (ng/ml)	3.85 $\pm$ 1.97	2.89 $\pm$ 1.32	2.99 $\pm$ 1.50	2.91 $\pm$ 1.40	2.88 $\pm$ 0.30	3.03 $\pm$ 1.45	3.01 $\pm$ 0.45	3.20 $\pm$ 0.80	3.55 $\pm$ 1.23
CEA (ng/ml)	2.77 $\pm$ 1.95	1.82 $\pm$ 0.98	1.94 $\pm$ 1.02	1.94 $\pm$ 1.04	1.85 $\pm$ 0.00	1.88 $\pm$ 1.00	1.76 $\pm$ 0.96	1.77 $\pm$ 0.90	2.12 $\pm$ 0.65

*Disease associated with risk of process of malignancy, **pre-malignancy.

3.2.3 CEA

As in the case of AFP, CEA values were also lower in the low-risk group ( $<$ 5.0 ng/ml) compared to health disease-free group (Table 1C). The mean value of CEA of all ages together in the low-risk group was 1.85 $\pm$ 1.03 ng/ml compared to 2.09 $\pm$ 1.28 ng/ml in the health diseases-free group, showing an inverse significant difference (Table 1C).

3.3 Concentration distribution of STK1p, AFP and CEA in different age groups

In this part of the study we extended the analysis of the low-risk groups by determining the distribution of the concentrations of the biomarkers to see if it could give further prognostic information. The low-risk groups were divided into three age groups (21–40, 41–60, $>$ 60 years) and compared to health disease-free persons. The health diseases-free group was not divided into ages, because of low number of persons ( $n=$ 428).

3.3.1 STK1p

In the health disease-free group the STK1p concentration was almost normal distributed between 0.1–0.6 pM (86.9%) with a limited tail up to 2.0 pM (13.6%) (Fig. 1A-a). No persons with STK1p above the cut-off value were found. The mean STK1p value was 0.38 $\pm$ 0.30 pM and the median 0.31 pM (Fig. 1A-a). In the low-risk group there were still persons with STK1p between 0.1–0.6 pM ( $\sim$ 53%), but the tail-persons with higher STK1p increased ( $\sim$ 43%) and even above 2.0 pM ( $\sim$ 4%) (Fig. 1A-b, A-c, A-d), increased the mean value to about 0.7 pM. No significantly differences in the concentration distribution among the ages was observed.

3.3.2 AFP

In the health disease-free group 98.8% of the AFP concentrations were within a normal distribution with a tail up to 10.0 ng/ml of 1.2% (Fig. 1B-a). No person above the cut-off value was found. The concentration distribution of the low-risk group was very similar to the health disease-free group, however, the mean values were lower (Fig. 1B-b, B-c, B-d). No significantly differences between the ages was found.

3.3.3 CEA

An almost normal distribution of the CEA concentration was found in the health disease-free group (97.5% below cut-off, and 2.5% above cut-off) (Fig. 1C-a). An almost normal distribution of the CEA concentration was also found in the low-risk group with increasing number of persons above the cut-off value by age, from 1.4% to 8.1% (Fig. 1C-b, C-c, C-d). The mean values were almost similar between the health disease-free group and the low-risk group.

3.4 Correlation between STK1p, AFP and CEA

In order to understand the relation between STK1p, AFP and CEA in serum of person in the low-risk group we determined the correlation between those biomarkers. Here we used the total number of persons, since there was no significantly difference between gender or ages.

There were no significantly correlations between STK1p and AFP, STK1p and CEA, or CEA and AFP (Fig. 2).

Table 3
STK1p and PSA values in low-risk persons with different types of prostate diseases including

Type of	Health	**Moderate/	*Prostate	*Prostate
prostate	disease-free	severe prolf.	calcification	cyst
STK1p (pM)	0.38 $\pm$ 0.30	0.64 $\pm$ 0.55	0.65 $\pm$ 0.64	0.69 $\pm$ 0.58
PSA (ng/ml)	1.52 $\pm$ 0.95	1.14 $\pm$ 0.68	0.97 $\pm$ 0.59	1.01 $\pm$ 0.65

**Pre-malignancy, *diseases associated with the risk process of malignancy.

3.5 Did biomarkers in low-risk group correlate to diseases?

The mean value of STK1p of the low-risk group was significantly higher compared to health disease-free group. The concentration distribution of STK1p also differed between low-risk group and health disease-free group with a shift to higher STK1p values. This were not seen for AFP and CEA. In the contrary, the mean AFP and CEA values were lower in the low-risk group compared to the health disease-free group and there was no significant increase in the number of low-risk persons towards higher AFP and CEA values. Thus, did persons with higher STK1p values (0.6–2.0 pM) compared to health disease-free persons (0.3–0.4 pM), but still not above the cut-off value ( $>$ 2.0 pM), indicate higher risk to develop pre-malignancies?

In Table 2 we show STK1p, AFP and CEA values of persons who had different liver diseases, including pre-malignancy and diseases associated with the risk process of malignancy. All of them show significantly higher STK1p values (0.6–1.0 pM) compared to the health disease-free group (0.3–0.4 pM), but not for AFP (health 3.9 ng/ml; low-risk 2.9–3.5 ng/ml) and CEA (health 2.0 ng/ml; low-risk 1.7–2.0 ng/ml).

We also investigated the STK1p and PSA values in the low-risk group (below cut-off) with different types of prostate diseases (Table 3). While the STK1p value was significantly higher in all prostate groups compared to the health disease-free group, the PSA values in persons with various prostate disease were even less compared to corresponding PSA values in the health disease-free group.

4. Discussion

In this study, we further analysed whether mean values and concentration distributions of STK1p, AFP, CEA and PSA among low-risk people (below cut-off values) were related to pre-malignant diseases associated with tumor process. We found that only STK1p correlated to people with pre-malignant diseases, but not AFP, CEA or PSA. This open up for the use of STK1p as a marker for early detection of risk to develop cancer in health screening.

Characterization of concentration distribution of STK1p has been done previously (review Skog et al. [12]) showing a shift from a normal distribution in the health disease-free group to a normal distribution with extensive tail of elevated STK1p values in the low-risk group (people below cut-off value of 2.0 pM). These results were based on three independent health screening studies, totally 38,121 persons [21, 22, 23] and summary by Skog et al. [12]. Here we extended these studies by investigated the correlation to gender and ages.

STK1p in the health disease-free group was almost normal distributed between 0.1–0.6 pM (86.9%) with a limited tail up to 2.0 pM (13.6%). The mean STK1p value was 0.38 $\pm$ 0.30 pM. On the other hand, in the health screening group at low-risk group (below the cut-off value of 2.0 pM) the number of persons in the concentration range of 0.1–0.6 pM decreased to 53%, while the number of persons in the range 0.6–2.0 pM increased to 43%, resulted in an elevated mean value of STK1p of 0.7 pM. The number of persons in the high-risk group ( $>$ 2.0 pM) also increased to 4.0%. In opposite to STK1p, no changes in the mean values and concentration distributions were found with AFP, CEA or PSA between health disease-free and low-risk groups (below cut-off values). AFP and CEA concentrations were normal distributed with almost no tails. Furthermore, there was no differences in the mean values and concentration distributions among gender and ages of STK1p, AFP and CEA. A likely explanation to the elevated STK1p values in the range of 0.6–2.0 pM is that: 1) STK1p is a proliferation marker; 2) pre-malignancies and diseases linked to development of tumor are related to proliferation; 3) a high frequency ( $\sim$ 50%) of people with pre-malignancies and diseases linked to development of tumor were found in the low-risk group [21, 22, 23].

Of particular interest in this study is the close correlation between the elevated STK1p values of the low-risk group (0.6–2.0 pM) and the presence of moderate/ severe type of pre-malignancies and diseases linked to development of tumor in liver and prostate. This type of correlation was, however, not found for AFP, CEA and PSA, supporting that STK1p is a more reliable biomarker for early detection of pre-malignancies and diseases linked to development of tumor compared to other tumor-related markers. Furthermore, since STK1p is a proliferation marker [8, 9, 10, 11, 12], it also indicates that these types of pre-malignancies found in the low-risk group are linked to abnormal cell proliferation. Although it is known from previous health screenings [20, 23], that persons with STK1p values above cut-off values ( $>$ 2.0 pM) have 3–5 times higher risk to develop malignancies compare to persons with STK1p values below the cut-off value [12], our results here show that even persons with STK1p values in the range of 0.6–2.0 pM have an increased risk to develop pre-malignancies. This is new observation. Thus, people with STK1p values between 0.6–2.0 pM should be a little concern and check up their health from time to time, however, not as frequent as those people with STK1p values above the cut-off value ( $>$ 2.0 pM).

A similar study on STK1p as presented here was performed by Cao et al. previously [25] using the same assay. The study included 14,960 Chinese participates (9,586 males and 5,374 females) from 20 to 79 years old, excluding people who had pre-cancer (moderate/severe type of hyperplasia of breast, prostate, gastrointestinal, cervix, liver cirrhosis, refractory anemia), or risk-diseases associated with tumor progression such as, liver disease, moderate/severe fatty liver, high risk for hepatitis B, abnormal liver function, obese, benign tumors (such as renal, thyroid), severe cardiac disease, using any medication that could affect STK1 levels such as exogenous hormone therapy, pregnant women, suffer from acute illness such as inflammation/virus infection within 4 weeks. The remaining persons consist of persons with minor proliferating /chronic/ non-tumor diseases that could not affect the STK1p levels. It was found that the STK1p levels decreased from 0.51 to 0.36 pM between 20–40 years, reaching a plateau value of 0.35 pM in late adulthood. The STK1p values were similar to the STK1p values we found among the health disease-free people above 41 years in our study (0.38 pM). Thus, the result from the health disease-free group and people with minor proliferating /chronic/ non-tumor diseases show similar STK1p values, significantly lower compare to people with pre-malignancies or risk diseases associated with tumor progression. Thus, the STK1p could distinguish between people of health disease-free/ minor proliferating /chronic/ non-tumor diseases and people with pre-malignancies or risk diseases associated with tumor progression.

Taking our results and the results of Cao et al. [25] together, it indicates that our STK1p assay is a reliable test below 2.0 pM for pre-malignancies or a variety of diseases associated with the risk of cancer progression later in life. It proved that the STK1p assay developed by us achieves high sensitivity and specificity and therefore should be useful in routine health screening. STK1p can be an alternative or complementary approach to discover the process of early malignant tumors.

According to previous reviews from Cigna [26] based on American Society of Clinical Oncology (ASCO), American Cancer Society (ACS), and National Cancer Institute (NCI), CEA and AFP were not recommended as a health screening tool due to low specificity. A likely explanation is that CEA and AFP, as many other bio-markers, are substances produced, not only by the tumor, but also by other type of cells in the body in response to pre- and malignancy. Tumor markers are found in normal cells and could show higher level compared to tumor cells. Furthermore, tumor markers may not be elevated in the presence of a tumor, especially in early stages of the disease, may not be specific to a particular type of tumor, and/or may be elevated by more than one type tumor. Despite these limitations, tumor-related markers are successfully combined with other diagnostic tools (e.g., laboratory tests, biopsy, radiological imaging) to confirm diagnosis/prognosis.

Opportunities for identifying high-risk individuals may be joined by application of receiver operation characteristic (ROC) statistical analysis. In a ROC test of STK1p between pre-operative malignant patients ( $n=$ 720) and normal health/a variety of mild disease, excluding pre-cancer individuals ( $n=$ 4,103) [23], the AUC value (area under the curve in the ROC analysis) was 0.96. In addition, the likelihood ( $+$ ) value was very high, too (236.5). At a STK1p risk cut-off value of 2.0 pM, the sensitivity and specificity were 0.80 and 0.99, respectively, showing low number of false positive cases (1/300). In an 11 years follow-up study [23] on people with elevated STK1p values ( $n=$ 170, $>$ 2.0 pM) and on randomly selected people with low STK1p ( $n=$ 6,354, $<$ 2.0 pM), it was shown that persons with elevated STK1p had a 3–5 times higher risk to develop malignancies compared to people with low STK1 values. Furthermore, the appearance of new tumors in the low-risk STK1p group was delayed by 11 years, while new tumors in the high-risk STK1p group appeared already at 6 years. Thus, development of a tumor is a slower process among people with low STK1p values compare to people with high STK1p values.

In summary, according to the present and previously published results [20, 23, 25] it is obvious that already at STK1p values below the recommended cut-off value of 2.0 pM, i.e. 0.6–2.0 pM, there is an increased risk to develop pre-malignancy and in the end malignancy. How to use the STK1p in routine health screening? We suggest that the STK1p concentration shall be combined with individuals who had pre-malignancy or diseases associated with the risk process of malignancy specifically. In addition, the individual STK1 value in health screening should be divided into two STK1p sub-groups: 1. Initial risk step of development of pre-malignancy or diseases associated with the risk process of malignancy of STK1p values of 0.6–2.0 pM, recommends health screening annual; 2. High risk step of development of malignancy or diseases associated with the risk process of malignancy of STK1p $>$ 2.0 pM, contact oncology doctor for early discovery of tumor and start treatment, increasing the chance to cure the cancer patient. The use of two-step risk assessment in health screening would increase the efficiency of discover people with risk to develop malignancy later in their life.

Footnotes

Acknowledgments

We acknowledge the Health Management Centre of PLA 180 Hospital, Quanzhou, China. We also appreciate the Sino-Swed Tongkang Biotech. Ltd., Shenzhen, China, for providing technical support.

References

World Health Statistics 2017, in: Monitoring Health for the SDGs, Sustainable Development Goals, ISBN 978-92-4-156548-6 © World Health Organization 2017, p. 31.

Zhang

Yan

Wan

Pang

Guo

Niu

Xiang

Zheng

and Xian

, Cancer burden in China from 2006 to 2010, Int J Clin Exp Pathol 8 (2015), 13323–13330.

Chen

W.O.

Zheng

R.S.

Zhang

S.W.

Zhao

Zeng

H.M.

and Zou

X.N.

, Report of cancer incidence and mortality in China, 2010, Ann Transl Med 2 (2014), 1–25.

Wang

Jiang

X.R.

Dong

Shen

Zhou

and Skog

, Serum TK1 is a more reliable marker than CEA and AFP for cancer screening in a study of 56286 people, Cancer Biomark 16 (2016), 529–536.

Stewart

B.W.

and Wild

Ch.P.

, World Cancer Report, 2014, ISBN 978-92-832-0429-9.

J.T.

, Circulating Tumor Markers of the New Millennium, AACC Press, New York, 2002, pp. 115–117, .

Gasparri

Wang

Skog

Galvani

and Eriksson

, Thymidine kinase 1 expression defines an activated G1 state of the cell cycle as revealed with site-specific antibodies and ArrayScan™ assays, Eur J Cell Biol 88 (2009), 779–785.

O.’Neill

K.L.

Buckwalter

M.R.

and Murray

B.K.

, Thymidine kinase: Diagnostic and prognostic potential, Expert Rev Mol Diagn 1 (2001), 428–433.

Topolcan

and Holubce

, Jr., The role of thymidine kinase in cancer diseases, Expert Opin Med Diagn 2 (2008), 129–141.

10.

Aufderklamm

Todenhofer

Gakis

Kruck

Hennenlotter

Stenzl

and Schwentner

, Thymidine kinase and cancer monitoring, Cancer Lett 316 (2012), 6–10.

11.

Zhou

and Skog

, The proliferation marker thymidine kinase 1 in clinical us, Mol Clin Oncol 1 (2013), 18–28.

12.

Skog

and Haghdoost

, in: Prevention and Early Detection of Human Tumo, LAP LAMBERT Academic Publishing, Berlin, Germany, Schaltungsdienst Lange O.H.G., Berlin, Germany, 2017, pp. 153–257.

13.

Guan

Sun

Zan

Zhou

Eriksson

Wen

and Skog

, Thymidine kinase 1 expression in atypical ductal hyperplasia significantly differs from usual ductal hyperplasia and ductal carcinoma in situ: A useful tool in tumor therapy management, Mol Med Rep 2 (2009), 923–929.

14.

Shi

Q.L.

Zhou

Eriksson

and Skog

, High thymidine kinase 1 (TK1) expression is a predictor of poor survival in patients with pT1 of lung adenocarcinoma, Tumor Biol 33 (2012), 475–483.

15.

Chen

Lin

Zheng

and Skog

, Nuclear TK1 expression is an independent prognostic factor for survival in pre-malignant and malignant lesions of the cervix, BMC Cancer 13 (2013), 2696–2702.

16.

Liu

Shi

Q.L.

Huang

P.L.

Zhou

H.H.

Z.F.

Eriksson

and Skog

, Thymidine kinase 1 is a better prognostic marker than Ki-67 for pT1 adenocarcinoma of the lung, Int J Clin Exp Med 8 (2014), 7120–2128.

17.

Wang

Liu

Zhou

Guo

Shi

Eriksson

Zhou

and Skog

, Thymidine kinase 1 expression in ovarian serous adenocarcinoma is superior to Ki-67: A new prognostic biomarker, Tumor Bio 39 (2017), 1393383185.

18.

Gronowitz

J.S.

Bergstrom

Nou

Pahlman

Brodin

Nilsson

and Kallander

C.F.

Clinical and serologic markers of stage and prognosis in small cell lung cancer, A multivariate analysis, Cancer 66 (1990), 722–732.

19.

Gronowitz

J.S.

Hagberg

Källander

C.F.

and Simonsson

, The use of serum deoxythymidine kinase as a prognostic marker, and in the monitoring of patients with non-Hodgkin’s lymphoma, British J Cancer 47 (1983), 487–495.

20.

Zou

Zhang

P.A.

H.X.

Wang

X.Q.

Zou

and Skog

, Concentration of thymidine kinase 1 in serum (S-TK1) is a more sensitive proliferation marker in human solid tumors than its activity, Oncol Rep 14 (2005), 1013–1019.

21.

Chen

Zhou

and Skog

, Serological thymidine kinase 1 (STK1) indicates an elevated risk for the development of malignant tumours, Anticancer Res 28 (2008), 3897–3907.

22.

Huang

Lin

Guo

Zhang

Yun

Liu

Zhou

and Skog

, Elevated serum thymidine kinase 1 predicts risk of pre/early cancerous progression, Asian-Pac J Cancer Prev 12 (2011), 497–505.

23.

Chen

Z.H.

Huang

S.O.

Wang

Yang

A.Z.

Wen

X.H.

Chen

Q.B.

Wang

Y.H.

Zhou

and Skog

, Serological thymidine kinase 1 is a biomarker for early detection of tumours – a health screening study on 35,365 people, using a sensitive chemiluminescent dot blot assay, Sensors (Basel) (2011), 11064–11080.

24.

Berman

J.J.

and Moore

G.W.

, in: Precancer: The Beginning and the End, Jones and Bartlett, eds, London, UK, 2010, pp. 118–126.

25.

Cao

Zhou

and Chen

Z.H.

, Standardized centile curves and reference intervals of serum thymidine kinase 1 levels in a normal chinese population using the LMS method genet, Test Mol Biomarkers 20 (2016), 445–450.

26.

Cigna, in: Tumor Markers for Cancer in Cigna Medical Coverage Policy, 2015, pp. 1–5.

Serological TK1 predict pre-cancer in routine health screenings of 56,178 people

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSION:

Keywords

1. Introduction

Table 1 The mean and median values of STK1, AFP and CEA with progressive ages

2.1 Participants

2.2 Methods

2.2.1 Discovery of tumor-related symptom

2.2.2 Serological assays

3. Results

3.2 Comparison of health disease-free and low-risk groups

3.2.1 STK1p

3.2.2 AFP

Table 2 STK1p, AFP and CEA values in the low-risk group with different types of liver pre-malignancy diseases

3.3 Concentration distribution of STK1p, AFP and CEA in different age groups

3.3.1 STK1p

3.3.2 AFP

3.3.3 CEA

3.4 Correlation between STK1p, AFP and CEA

Table 3 STK1p and PSA values in low-risk persons with different types of prostate diseases including

4. Discussion

Footnotes

Acknowledgments

References

Table 1
The mean and median values of STK1, AFP and CEA with progressive ages

Table 2
STK1p, AFP and CEA values in the low-risk group with different types of liver pre-malignancy diseases

Table 3
STK1p and PSA values in low-risk persons with different types of prostate diseases including