Opportunities to Avoid Invasive Cancer by Diagnosis and Interception of Preneoplastic Lesions and Cancer Risk Conditions

Abstract

Background:

Preneoplastic lesions (PNL) represent a state intermediate between normal cells and invasive cancer (IC). Individuals diagnosed with PNL are at increased risk of developing IC and represent a unique risk group. We estimated the proportion of IC that could be intercepted at multiple anatomic sites in individuals with PNL.

Methods:

Using a literature survey of the prevalence, latency, and probability of progression of PNL to IC, we undertook a sensitivity analysis across proportions of IC from 5%-95% that could be intercepted in a cohort of 100,000 individuals that would be avoided if PNL interception was successful. To calculate the number of avoidable IC, we used data that represent population-based, high-risk, unselected, and selected PNL case series, and calculated the sensitivities and specificities of avoidable ICs.

Results:

Substantial invasive cancer reductions could be achieved by interception of PNL that precede pancreas, gastric, bladder, prostate, breast, colorectal, and skin cancers. Intermediate reduction in cancer incidence was observed for lung, and oral cancers. Limited impact on the global burden of cancer with interception of PNL are likely for cervical, liver, and esophageal cancer as well as myeloma and leukemia. Due to the estimates representing varying populations, impacts of a specific PNL or cancer cannot be compared with another.

Conclusion:

We identified substantial variation in the impact that interception of PNL would have on the development of IC.

Impact:

These results may guide research and implementation of studies intended to maximize the number of IC that can be avoided by interception of PNL.

INTRODUCTION

Preneoplastic lesions (PNL) represent morphologically identifiable cellular and tissue traits that have the potential to progress to invasive cancer. PNL have been referred to by a variety of names including precancers, precancerous lesions, precursor lesions, premalignant lesions, premalignancies, incipient cancers, intraepithelial neoplasia, or preinvasive cancers. In 2004, an NCI panel defined PNL to have the following characteristics: PNL are associated with an increased risk of invasive cancer; when PNL progress to invasive cancer, the tumor arises from cells within the PNL; PNL differ with respect to molecular or morphological characteristics from the normal tissue from which it arises; PNL differ molecularly and morphologically from the cancer into which it develops; and the PNL itself can be detected and diagnosed(1). When these criteria are fulfilled, PNL present an opportunity to identify those at elevated risk of invasive cancer and minimize cancer-associated mortality through early detection and interception(2,3). This definition has been more recently expanded to include the biological and molecular underpinnings around which PNL are defined, detected, and intercepted(4).

For example, preneoplastic squamous lesions of the bladder have a prevalence of 10%. Among those individuals with the lesion, there is between a 30-78% probability of progressing to invasive bladder cancer within the individuals lifetime(5-7). Similarly, preneoplastic gastric intestinal metaplasia of the stomach can have a prevalence of up to 19% and is associated with up to a 42% probability of progression to invasive stomach cancer(8-10). These and other data suggest that there is a substantial risk of invasive cancer in those with a PNL. Morbidity and mortality can be avoided by diagnosing PNL and intercepting their progression to invasive cancer.

To date there have been limited data that demonstrate the potential reduction of invasive cancer through interception of PNL. To understand this impact at a population level, we estimated the hypothetical number of invasive cancers that could be avoided in the United States population by the successful interception of PNL.

MATERIALS AND METHODS

Prevalence, Latency, and Progression

We searched the published literature to obtain estimates of prevalence, latency, and probability of progression of PNL to invasive cancer. We identified the following PNL that have been associated with invasive cancers: squamous lesions for invasive bladder cancer, colon polyps or advanced adenoma for invasive colorectal cancer, non-alcoholic fatty liver disease or non-alcoholic steatohepatitis or cirrhosis for invasive liver cancer, oral leukoplakia for invasive oral cancer, serous tubal intraepithelial carcinoma for ovarian cancer, pancreatic intraepithelial neoplasia for invasive pancreatic cancer, prostatic intraepithelial neoplasia for invasive prostate cancer, and actinic keratosis for invasive skin squamous cell cancer (Table 1, Supplementary Table 1, and (Supplementary Table 2). For the purpose of this study, it is important to clarify that we used the term PNL to encompass conditions and/or lesions of abnormal cells that confer an enhanced risk of potentially developing into benign or malignant tumors. These include neoplastic as well as non-neoplastic states (e.g., cirrhosis). Estimates of PNL prevalence, latency, and percent progression were obtained from a search of the literature to represent a range of values from which an avoidable fraction could be computed (Supplementary Table 2). These values were obtained from clinical and population reports as there are no single cited studies of progression of PNL to invasive cancer. The descriptions of the individual populations included in each study are provided in Supplementary Table 2 for reference.

Table 1: Values used to estimate Proportion of Cancers Avoidable in a Population of 100,000 individuals at risk.

Invasive Cancer Site	Preneoplastic Lesion	Prevalence	Individuals with Lesion from a Cohort of 100,000	Percent Progression	Latency (Years)	PPYa
Breast(67,68)	Atypical Hyperplasia	10.00%	10,000	20.00%	8.3	2.41%
Bladder(5-7)	Squamous lesions (Bladder Carcinoma in situ)	10.00%	10,000	31.00%	29	1.07%
	Squamous lesions (Bladder Carcinoma in situ)	10.00%	10,000	78.00%	29	2.69%
Cervix(69-71)	Cervical Intraepithelial Neoplasia (CIN 1)	2.30%	2,300	6.00%	12	0.50%
	Cervical Intraepithelial Neoplasia (CIN 1)	2.30%	2,300	12.00%	12	1.00%
	Cervical Intraepithelial Neoplasia (CIN 2)	0.50%	500	6.00%	12	0.50%
	Cervical Intraepithelial Neoplasia (CIN 2)	0.50%	500	12.00%	12	1.00%
	Cervical Intraepithelial Neoplasia (CIN 3)	0.40%	400	6.00%	12	0.50%
	Cervical Intraepithelial Neoplasia (CIN 3)	0.40%	400	12.00%	12	1.00%
Colorectum(72,73)	Colon Polyps Male	24.00%	24,000	25.00%	5	5.00%
	Colon Polyps Male	24.00%	24,000	25.00%	15	1.67%
	Colon Adenoma Male	13.00%	13,000	25.00%	5	5.00%
	Colon Adenoma Male	13.00%	13,000	25.00%	15	1.67%
	Advance Adenoma Male	3.00%	3,000	25.00%	5	5.00%
	Advance Adenoma Male	3.00%	3,000	25.00%	15	1.67%
	Colon Polyps Female	24.00%	24,000	15.00%	5	3.00%
	Colon Polyps Female	24.00%	24,000	15.00%	15	1.00%
	Colon Adenoma Female	13.00%	13,000	15.00%	5	3.00%
	Colon Adenoma Female	13.00%	13,000	15.00%	15	1.00%
	Advance Adenoma Female	3.00%	3,000	15.00%	5	3.00%
	Advance Adenoma Female	3.00%	3,000	15.00%	15	1.00%
Esophagus(74,75)	Barrett's Esophagus (Prevalence 1)	0.80%	800	0.33%	20	0.02%
	Barrett's Esophagus (Prevalence 2)	28.00%	28,000	0.33%	20	0.02%
Liver(76,77)	Non-Alcoholic Fatty Liver Disease	4.00%	4,000	15.00%	30	0.50%
	Non-Alcoholic Fatty Liver Disease	4.00%	4,000	15.00%	40	0.38%
	Non-Alcoholic Steatohepatitis (Prevalence 1)	1.00%	1,000	15.00%	30	0.50%
	Non-Alcoholic Steatohepatitis (Prevalence 1)	1.00%	1,000	15.00%	40	0.38%
	Non-Alcoholic Steatohepatitis (Prevalence 2)	6.00%	6,000	15.00%	30	0.50%
	Non-Alcoholic Steatohepatitis (Prevalence 2)	6.00%	6,000	15.00%	40	0.38%
	Cirrhosis (Prevalence 1)	1.00%	1,000	15.00%	30	0.50%
	Cirrhosis (Prevalence 1)	1.00%	1,000	15.00%	40	0.38%
	Cirrhosis (Prevalence 2)	6.00%	6,000	15.00%	30	0.50%
	Cirrhosis (Prevalence 2)	6.00%	6,000	15.00%	40	0.38%
Leukemia(78-80)	Clonal Hematopoiesis of Indeterminant Potential (Prevalence 1)	0.50%	500	0.50%	10	0.05%
	Clonal Hematopoiesis of Indeterminant Potential (Prevalence 1)	0.50%	500	1.00%	10	0.10%
	Clonal Hematopoiesis of Indeterminant Potential (Prevalence 2)	1.00%	1,000	0.50%	10	0.05%
	Clonal Hematopoiesis of Indeterminant Potential (Prevalence 2)	1.00%	1,000	1.00%	10	0.10%
Lung(81,82)	Lung Nodules	9.40%	9,400	2.60%	2	1.30%
Myeloma(13,83-86)	Monoclonal Gammopathies of Undetermined Significance	1.00%	1,000	1.00%	5	0.20%
	Monoclonal Gammopathies of Undetermined Significance	3.00%	3,000	1.00%	5	0.20%
Oral(87-89)	Oral Leukoplakia (Prevalence 1)	1.50%	1,500	0.00%	3	0.00%
	Oral Leukoplakia (Prevalence 1)	1.50%	1,500	0.00%	5	0.00%
	Oral Leukoplakia (Prevalence 1)	1.50%	1,500	36.00%	3	12.00%
	Oral Leukoplakia (Prevalence 1)	1.50%	1,500	36.00%	5	7.20%
	Oral Leukoplakia (Prevalence 2)	4.30%	4,300	0.00%	3	0.00%
	Oral Leukoplakia (Prevalence 2)	4.30%	4,300	0.00%	5	0.00%
	Oral Leukoplakia (Prevalence 2)	4.30%	4,300	36.00%	3	12.00%
	Oral Leukoplakia (Prevalence 2)	4.30%	4,300	36.00%	5	7.20%
Ovarian(90-92)	Serous Tubal Intraepithelial Carcinoma (Prevalence 1)	3.50%	3,500	35.00%	6.5	5.38%
	Serous Tubal Intraepithelial Carcinoma (Prevalence 1)	3.50%	3,500	50.00%	6.5	7.69%
	Serous Tubal Intraepithelial Carcinoma (Prevalence 2)	5.60%	5,600	35.00%	6.5	5.38%
	Serous Tubal Intraepithelial Carcinoma (Prevalence 2)	5.60%	5,600	50.00%	6.5	7.69%
Pancreas(93-98)	Pancreatic Intraepithelial Neoplasia (PanIN1)	2.00%	2,000	1.00%	5	0.20%
	Pancreatic Intraepithelial Neoplasia (PanIN1)	2.00%	2,000	10.00%	5	2.00%
	Pancreatic Intraepithelial Neoplasia (PanIN1)	2.00%	2,000	15.00%	5	3.00%
	Pancreatic Intraepithelial Neoplasia (PanIN1)	13.00%	13,000	1.00%	5	0.20%
	Pancreatic Intraepithelial Neoplasia (PanIN1)	13.00%	13,000	10.00%	5	2.00%
	Pancreatic Intraepithelial Neoplasia (PanIN1)	13.00%	13,000	15.00%	5	3.00%
	Pancreatic Intraepithelial Neoplasia (PanIN1)	43.00%	43,000	1.00%	5	0.20%
	Pancreatic Intraepithelial Neoplasia (PanIN1)	43.00%	43,000	10.00%	5	2.00%
	Pancreatic Intraepithelial Neoplasia (PanIN1)	43.00%	43,000	15.00%	5	3.00%
	Pancreatic Intraepithelial Neoplasia (PanIN2)	13.00%	13,000	10.00%	5	2.00%
	Pancreatic Intraepithelial Neoplasia (PanIN2)	13.00%	13,000	15.00%	5	3.00%
	Pancreatic Intraepithelial Neoplasia (PanIN3)	3.00%	3,000	10.00%	5	2.00%
	Pancreatic Intraepithelial Neoplasia (PanIN3)	3.00%	3,000	15.00%	5	3.00%
Prostate(63,99)	Prostatic Intraepithelial Neoplasia (PIN) (Prevalence 1)	4.00%	4,000	24.00%	10	2.40%
	Prostatic Intraepithelial Neoplasia (PIN) (Prevalence 2)	16.00%	16,000	24.00%	10	2.40%
Skin Melanocyte(100-102)	Dysplastic Nevus (Prevalence 1)	2.00%	2,000	3.30%	20	0.17%
	Dysplastic Nevus (Prevalence 2)	8.00%	8,000	3.30%	20	0.17%
Squamous Cell (Skin) (103-105)	Actinic Keratosis Male	49.00%	49,000	10.00%	25	0.40%
	Actinic Keratosis Male	49.00%	49,000	10.00%	30	0.33%
	Actinic Keratosis Female	28.00%	28,000	10.00%	25	0.40%
	Actinic Keratosis Female	28.00%	28,000	10.00%	30	0.33%
Stomach(8-10,106,107)	Gastric Intestinal Metaplasia	2.50%	2,500	0.25%	4	0.06%
	Gastric Intestinal Metaplasia	2.50%	2,500	42.00%	4	10.50%
	Gastric Intestinal Metaplasia	4.80%	4,800	0.25%	4	0.06%
	Gastric Intestinal Metaplasia	4.80%	4,800	42.00%	4	10.50%
	Gastric Intestinal Metaplasia	19.00%	19,000	0.25%	4	0.06%
	Gastric Intestinal Metaplasia	19.00%	19,000	42.00%	4	10.50%

^aPPY = PP/L where PP is the percent progressing and L is the latency years

Statistical Analysis

To estimate the number of invasive cancer cases that could be avoided if interception of a PNL before it progressed to invasive cancer case was successful, we assumed a hypothetical sample of 100,000 individuals. To estimate the number of individuals with a particular preneoplastic lesion, we multiplied 100,000 by the prevalence value obtained from the literature (Table 1). The percent of PNL that could progress to invasive cancers per year were calculated based on values of L (latency in years), PP (percent of PNL that progress to invasive cancer), and PPY (percent progression per year) computed as PPY = PP/L. Values for L and PP were obtained from the literature (Table 1). The proportion of PNL that could be intercepted before progression to invasive cancer was calculated in a range of 5-95%. Finally, we calculated the positive predictive value (PPV) and negative predictive value (NPV) using the prevalence values in Table 1 and assuming a range of sensitivity and specificity values for each PNL, though these are not tool or intervention specific and provide a generic representation of the PPV and NPV interception calculation.

We assumed that in our hypothetical sample of 100,000 individuals, there were no competing risks and did take into consideration follow-up or censoring. Furthermore, we assumed that the prevalence, latency, and percent progression were held constant within each PNL regardless of case characteristics or other risk factors. We also assumed that there is an intervention available to prevent a percentage of the progression of these lesions to invasive cancer to highlight the potential value of identifying and testing interception strategies in these patients. When relevant data were available, we also computed the above-mentioned values stratified by sex and/or race.

Data Availability

Publicly available data generated by others were used by the authors and then further data were generated by the authors and included in the article.

RESULTS

Figures 1-5 depict the number of invasive cancer cases from a population of 100,000 individuals that could be avoided assuming PNL for the corresponding percent progression, prevalence, and latency value. The largest number of cancers that could be avoided are 1,226 invasive pancreatic cancer cases if 95% of pancreatic intraepithelial neoplasia (PanIN1) lesions were identified; with a 15% progression, 43% prevalence, and a 5-year latency period (Figure 1). Furthermore, if 95% of squamous lesions were identified then 256 invasive bladder cancer cases could be avoided. (Figure 1).

Figure 1: Estimates of the Number of Invasive Cancer Cases that Could be Avoided if Preneoplastic Lesion Progression Was Successfully Intercepted in General and High-Risk Populations: Pancreatic and Bladder.

Modeled estimates of the number of invasive pancreatic and bladder cancer cases that could be avoided if 5% to 95% of preneoplastic lesions were identified and intercepted before progression. Lesions modeled include pancreatic intraepithelial neoplasia (PanIN1, PanIN2, PanIN3) and squamous lesions associated with bladder carcinoma in situ. Estimates are shown across a range of lesion-specific assumptions for prevalence, percent progression, latency (years), and percent progression per year (PPY).

Figure 5: Estimates of the Number of Invasive Cancer Cases that Could be Avoided if Preneoplastic Lesion Progression Was Successfully Intercepted in General and High-Risk Populations: Hematological Malignancies.

Modeled estimates of the number of invasive hematological malignancy cases that could be avoided if 5% to 95% of preneoplastic lesions were identified and intercepted before progression. Lesions modeled include clonal hematopoiesis of indeterminant potential and monoclonal gammopathies of undetermined significance, stratified by race (White and Black). Estimates are shown across a range of lesion-specific assumptions for prevalence, percent progression, latency (years), and percent progression per year (PPY).

Additionally, there is potential to avoid a higher number of invasive cancer cases by identifying other PNLs of the digestive regions (Figure 2). For example, if 95% of gastric intestinal metaplasia lesions were identified; with a 42% progression, 19% prevalence, and a 4-year latency period; then 998 potential invasive stomach cancer cases could be avoided (Figure 2). Similarly, if 95% of oral leukoplakia lesions were identified; with a 36% progression, 4.3% prevalence, and a 3-year latency period; then 490 potential invasive oral cancer cases could be avoided (Figure 3). Among PNLs of the genitourinary region, if 95% of prostatic intraepithelial neoplasia and atypical hyperplasia are identified then 365 invasive prostate cancer cases and 229 invasive breast cancer cases could be avoided, respectively (Figure 4).

Figure 2: Estimates of the Number of Invasive Cancer Cases that Could be Avoided if Preneoplastic Lesion Progression Was Successfully Intercepted in General and High-Risk Populations: Digestive Cancers.

Modeled estimates of the number of invasive digestive cancer cases that could be avoided if 5% to 95% of preneoplastic lesions were identified and intercepted before progression. Lesions modeled include non-alcoholic fatty liver disease, non-alcoholic steatohepatitis, cirrhosis, and gastric intestinal metaplasia. Estimates are shown across a range of lesion-specific assumptions for prevalence, percent progression, latency (years), and percent progression per year (PPY).

Figure 3: Estimates of the Number of Invasive Cancer Cases that Could be Avoided if Preneoplastic Lesion Progression Was Successfully Intercepted in General and High-Risk Populations: Upper Airway Cancers.

Modeled estimates of the number of invasive upper airway cancer cases that could be avoided if 5% to 95% of preneoplastic lesions were identified and intercepted before progression. Lesions modeled include Barrett’s esophagus, lung nodules, and oral leukoplakia. Estimates are shown across a range of lesion-specific assumptions for prevalence, percent progression, latency (years), and percent progression per year (PPY).

Figure 4: Estimates of the Number of Invasive Cancer Cases that Could be Avoided if Preneoplastic Lesion Progression Was Successfully Intercepted in General and High-Risk Populations: Hormonal/Reproductive Cancers.

Modeled estimates of the number of invasive hormonal or reproductive cancer cases that could be avoided if 5% to 95% of preneoplastic lesions were identified and intercepted before progression. Lesions modeled include cervical intraepithelial neoplasia (CIN 1, CIN 2, CIN 3), prostatic intraepithelial neoplasia, atypical hyperplasia, and serous tubal intraepithelial carcinoma. Estimates are shown across a range of lesion-specific assumptions for prevalence, percent progression, latency (years), and percent progression per year (PPY).

The impact of interception may vary by race. Figure 5 shows the difference among PNL of monoclonal gammopathies of undetermined significance for myeloma stratified by race. The prevalence of monoclonal gammopathies of undetermined significance among Black and White individuals is 3.70% and 2.30%, respectively. Both populations have a percent progression of 1% and latency period of 5-years for this preneoplastic lesion (Figure 5). Due to the higher prevalence among Black individuals, if 95% of these lesions are identified earlier, then 7 invasive myeloma cancer cases (per 100,000 Black individuals) could be avoided among Black individuals compared to 4 cases (per 100,000 White individuals) among White individuals (Figure 5).

Results were also stratified by sex for which values were available in the literature. For PNL of colon adenoma and advanced adenoma, males had a higher percent progression (25%) to invasive colorectum cancer compared to females (15%). With a higher percent progression, there was a larger number of preventable invasive cancer cases regardless of if 5% or 95% of PNL were detected earlier, among males compared to females (Supplementary Figure 1). A similar trend was present for actinic keratosis which has a prevalence of 49% among males and 28% among females with relation to invasive skin squamous cell cancer (Supplementary Figure 2).

DISCUSSION

This analysis suggests that a substantial number of invasive cancer cases could be avoided through successful detection and interception of PNL. Our PPV and NPV estimates confirm that avoidability of invasive cancer among PNL is highly dependent on PNL prevalence and percent progression per year (Supplementary Table 3 and Supplementary Table 4). Thus, it is possible to prioritize certain PNL as targets for detection and interception of invasive cancers to maximize impact on population cancer burden.

Our study is one of the first to estimate the quantitative impact of detecting and intercepting PNL before a diagnosis of invasive cancer. It is also one of the first studies to synthesize data from several analyses of PNL and their methods for identification and intervention. However, there are a few limitations in our study such as the various assumptions which were made to quantity the differences between races and calculate the PPV and NPV values. These assumptions were made to show hypothetical impact and need to enhance cancer early detection methods, due to the limiting data available for prevalence, latency, and percent progression in the literature. Additionally, we assumed a hypothetical population representing a distribution of individuals with differing characteristics. This approach does not include death from other causes or a survival model-based approach which would allow for an evolving baseline cohort at risk over time, thus not taking into consideration competing risks, but this could also potentially require further assumptions. We have chosen representative values taken from the literature. Many of these estimates do not represent the general population but are representative of high-risk individuals or highly selected case series. We do not consider this a limitation of the present analysis because they represent populations who have been identified as having PNL and are therefore the groups in whom interventions may be applied. Furthermore, interventions are also often unlikely to be broadly applied to the general population but rather to those individuals who are already at a higher risk, making these estimates potentially more relevant to real-world decision-making. From a screening and prevention perspective, it is these individuals who would most benefit from stratified risk assessment and subsequent intervention. This approach is also consistent with prior studies of precancerous lesions, which commonly rely on enriched or symptomatic cohorts to assess potential clinical impact. Therefore, though some of these estimates are from high-risk populations resulting in values which may be elevated compared to the average population, our study shows the enhanced potential for impact in those populations. Lastly, this is also beneficial for the long term because understanding lesion behavior in high-risk populations provides a critical foundation for translating findings to broader populations.

Variation in PNL Risk and Progression

When stratified by race for monoclonal gammopathies of undetermined significance, the literature indicates that Black individuals have a 2- to 3-fold higher prevalence for this specific preneoplastic lesion(11,12). Though PNL of monoclonal gammopathies of undetermined significance are already rare, if 50% of the cases are detected earlier, then we can potentially avoid the development of 4 cases of symptomatic myeloma per 100,000 Black individuals compared to 2 cases of symptomatic myeloma per 100,000 White individuals(12-14). This highlights the need to consider race when evaluating the burden from PNL and cancer, as they could affect a specific race significantly more than another(15,16). By doing so, we can hopefully reduce the burden of symptomatic cancer based on racial differences if caught earlier.

When stratified by sex, our results show that we can avoid several cases for both males and females with PNL for invasive colorectum and skin squamous cell cancer if detected early. For males, if 50% of PNLs are detected before progressing to invasive colorectum cancer, we can avoid up to 600 cases of colon polyps, 325 cases of colon adenoma, and 108 cases of advanced adenoma per 100,000 males. Similarly, we can avoid up to 360 cases of colon polyps, 195 cases of colon adenoma, and 45 cases of advance adenoma per 100,000 females. Similar results were found for actinic keratosis if 50% were detected early with up 98 cases per 100,000 males and up to 56 cases per 100,000 females.

Though data on prevalence and percent progression are limited on genetic markers for PNL, literature shows that PNL vary in severity due to their genetics, which could affect the progression from PNL to invasive cancer. For example, gastric intestinal metaplasia exhibits a large variation in the percent of PNL that progress to stomach cancer (0.25-42.00%). This then impacts the number of potential invasive cancer cases which can be detected earlier, preventing up to 525 cases from progressing to invasive cancer if (for example) 50% are intercepted. This can be further impacted by the type of molecular marker characterizing the intestinal metaplasia, such as if it’s a complete marker (MUC2) or incomplete marker (MUC2, MUC5AC, MUC6), which can then result in the formation of various subtypes of tumors and invasive gastric cancer(17,18). Similarly, the probability of the lung nodules progressing is affected by the presence or loss of heterozygosity on chromosome 3 and changes to TP53, both severing as molecular markers that may hallmark changes from PNL to invasive cancer(19). Similarly, MGUS is a common condition, and most MGUS do not progress to malignant myeloma(20). Thus, risk-adaptive approaches that identify the individuals at highest risk of malignant myeloma based on risk factors, somatic molecular signatures, and other features may be more appropriate than a blanket approach to intervention on all MGUS.

Benefits of PNL Detection and Interception

The economic burden from cancer care in the US has increased by 66% from $125 billion to $207 billion in 2020. Strategies to mitigate the economic burden of cancer are needed. The ability to detect potential cancer cases earlier may lead to not only more effective treatment, but also help reduce the cost of cancer(21). For example, McGarvey et al found that for a woman diagnosed with stage IV ovarian cancer, the absolute mean cost is $227,899, compared to if she had been diagnosed at stage I ovarian cancer, the absolute mean cost would be $73,428. Similarly, for colorectal cancer, the cost if diagnosed at stage IV is $271,216 compared to stage I is $129,645. The average cost for 6 months of chemotherapy is almost $27,000 and these costs are expected to increase by 34% by the year 2030(22-24). Similarly, results from a 2017 study show that for Medicare patients, common chemotherapies cost between $7,500-$25,000(25). On a national level, literature estimates that the United States as a nation would save around $26 billion a year if cancers were diagnosed earlier(26,27). Furthermore, in 2010 the economic cost of cancer through healthcare expenses and loss of productivity was estimated at $1.16 trillion for the United States and globally, this cost is expected to increase to $25.2 trillion by 2050(28,29).

These costs are estimated for invasive cancer. We therefore infer that interception of PNL before progression to invasive cancer, the economic burden would also be significantly reduced(30). The cost of resecting actinic keratosis has been around $200, depending on the type of procedure such as destructive procedure to remove the lesion or prescription of a topical medication(31,32). A recent study has quoted the average cost for treatment actinic keratosis to be $17.56 and for other skin related benign lesions to be $18.62, both per total body skin examination(33). In comparison, cutaneous squamous cell carcinoma in situ (CIS) incurs costs of approximately $463 for treatment with curettage and electrodessication, and up to $5,000 per lesion for Mohs surgery(34). Whereas for treatment of melanoma, these costs have continued to increase throughout the years, with treatment costing a mean of $33,347 per patient compared to $47,886 between 2007-2012 and 2018-2019, respectively(35). The costs of treatment between 2007-2012 for melanoma were $23,491; $46,511; and $47,739 for stages II-IV, respectively(35). Compared to 2018-2019, during which the costs increased to $27,035; $67,108; and $117,450 for stages II-IV, respectively, and these costs are expected to continue rising(35). Similar trends hold for other PNLs and their respective cancers, however by intercepting and resecting at this stage, there is significant economic benefit both at the individual and national level.

Along with reducing the burden on the national and global economy, individuals could also continue to work and support their families, reducing their own personal economic burden from the disease. A study by Altice et al found that cancer survivors experienced a loss of up to $8,236 in mean annual productivity. In addition, 12-62% of survivors reported being in debt due to treatment costs and 4-45% could not comply to recommended medications because of costs(36). Overall, by detecting PNL before they progress to invasive cancer, we can significantly reduce the individual, national, and global economic burden from cancer.

The psychological impacts of being diagnosed with PNL have been minimally studied. Some individuals experience the same level of stress with PNL that they may with invasive cancer(37,38). Reasons for this include that those individuals, regardless of if they have a preneoplastic lesion or invasive cancer, still feel isolated due to increased uncertainty and limited social support(37,38). Consequently, some previous research shows that eliminating the word “cancer” from the diagnosis, highly impacts the choices that individuals with low-risk neoplasms of certain cancers (such as thyroid, breast, melanoma, lung, and prostate) make particularly pertaining to those of overtreatment(39-44). Taking this a step further by detecting lesions at the precancerous stage, we hypothesize that there will be a reduction in the amount of psychological stress, both mental and emotional, a patient experiences compared to if they were diagnosed with invasive cancer. Though there is a lack of research on this, individuals with PNL may experience lower levels of psychological stress due to the possibility of being able to fully remove the lesion and prevent future invasive cancer and stress due to reoccurrence which cancer survivors often experience.

There is also the potential to prevent broader health complications by the early detection of some of these PNL’s. For example, clonal hematopoiesis of indeterminant potential can result in inflammation which have downstream effects by accelerating the osteo-hematopoietic niche and serving as a risk factor for bone and joint disorders such as osteoporosis and osteoarthritis(45,46). Clonal hematopoiesis of indeterminant potential has also been associated with adverse outcomes for other diseases including cardiovascular disease, type 2 diabetes, autoimmune disorders, gout, and chronic liver disease(45,47-52). In comparison to other PNLs, though the prevalence of percent progression of clonal hematopoiesis of indeterminant potential and some other PNLs is relatively low, early detection of them can have benefit beyond cancer detection and interception.

Potential Interceptions

With early detection of PNL, there is also opportunity for interception before progression to invasive cancer. By intervening, we can also reduce the burden of cancer on both the individual and global scale. Interception involves the early identification and diagnosis of the PNL, followed by either removal or treatment of the PNL, and, if required, continued monitoring. From Supplementary Table 1, there is possibility for various types of interceptions at the precancerous stage. Interception methods can vary from cancer prophylaxis, to cancer interception, cancer mitigation, precision cancer prevention, and chemoprevention(53). Cancer prophylaxis encompasses actions that can be taken early to avoid the transition of normal cells to potential cancer precursors and these actions are beneficial for those individuals for whom this transition process has yet to begun(53). These methods include consuming a healthy balanced diet which is rich in fruits and vegetables, frequent exercise, smoking cessation, alcohol restriction, restricted carcinogen exposures, vaccination, administration of exogenous agents such as aspirin, enhanced health policy changes, among others(53,54). Cancer interception methods are targeted for individuals who may have a PNL and are progressing toward a potential cancer or those with elevated cancer risk from family history, genetics, or previous exposures(53). Interception includes cancer screening, surgical prevention, excision of the PNL, as well as others; all of which can delay the onset of invasive cancer from the PNL(53,54). Furthermore, cancer mitigation techniques can be implied for those individuals with a PNL which may not be clinically evident and may be asymptomatic(53). These include molecular testing, genetic marker monitoring, imaging, and multicancer detection tests(53,54). While precision cancer prevention techniques take this one step further and target specific populations and incorporate germline genetic variation assays to further mitigate biological, genetic, demographic related risk(53). Lastly, chemoprevention methods; which can be useful for cancer prophylaxis, cancer interception, cancer mitigation, and precision cancer prevention; use substance and targeted drug therapeutics to prevent or delay carcinogenesis both at the PNL and invasive cancer level(53,54). For example, the administration of tamoxifen as a chemoprevention agent to reduce risk of invasive breast cancer(53). However, though there are several methods to prevent progression from PNL to invasive cancer, it is important to note that no such perfect interventions exist, and intervention will differ by cancer type, resulting in varying efficacies. Thus, the values provided in Supplementary Table 1 show the impacts from a hypothetical interception tool with the corresponding effect on invasive cancer avoidance described in our study.

Harms of PNL Detection and Interception

While it is clear from our analysis that there are benefits from avoidance of invasive cancer by detecting and intercepting PNL, there are also possible harms. Research has already shown that surgery or intervention of primary tumors can sometimes disrupt the natural history progression of cancer and lead to increased rates of metastasis, tumor seeding, as well as tumor cell proliferation, differentiation, differentiation, and angiogenesis(55-58). Similarly, with the potential for cancer early detection, there is potential for overdiagnosis and overtreatment of PNL, all of which can impact the natural history of cancer and change the risk of a future tumor. Overdiagnosis has been defined as the diagnosis of nonprogressive cancers, those which are growing but at a pace that is so slow that the individual will most likely die from another cause than before the cancer can show its symptoms(59). For example, studies have already shown evidence of overdiagnosis of indolent prostate cancer with prostate-specific antigen (PSA) testing and with the rise of enhanced early cancer detection technology, there will be an increase in the detection of these types of non-progressive cancers leading to unnecessary interventions such as surgery or radiation(59-61). Similar implications of overdiagnosis of indolent prostate cancer by detecting prostatic intraepithelial neoplasia lesions or high-grade prostatic intraepithelial neoplasia lesions, a precursor of prostate cancer, may also result in additional biopsies or overtreatment without demonstrated clinical benefit(62,63). Diagnosis of an indolent prostate cancer can result in increased risk of complications from treatments such as infection, erectile dysfunction, and blood loss(61). Along with physical effects, there is a higher probability of anxiety and depression resulting from the stress of a cancer diagnosis(61). Additionally, there is also the risk of increased morbidity from these interventions. Though surgery is one of the main curative treatments for pancreatic intraepithelial neoplasia, the risk of post operative morbidity can potentially be up to 50%(64,65). This highlights that though these early detection practices have several benefits, there must be balance when generating a knowledge base and translation of that knowledge base into clinical practice and policy. Though detection is important, the actionability of the information is the most important aspect.

Conclusions

By detecting PNL before they progress to invasive cancer, it may be possible to mitigate the burden of invasive cancer on individuals and the population. While these benefits are obvious, the development and implementation of PNL detection and interception must carefully consider the harms as well as benefits of these approaches. Such efforts should also consider the broader population in whom these lesions are commonly detected despite the absence of known elevated risk. Even if the individual benefit may be modest in these cases, the increased prevalence of incidental findings in lower-risk individuals may result in meaningful cumulative impact at the population level(66). Future research to better understand the distribution and determinants of PNL development, prevalence, and progression as well as interception opportunities is necessary to fulfill the potential of PNL to impact favorably on cancer interception and early detection.