Patent Application
20050282146
The sentinel lymph node (SLN) biopsy procedure has been validated by numerous studies and found to be accurate for assessing regional lymph node involvement (Giuliano et al., 1997; Albertini et al., 1996; Veronesi et al., 1997; O'Hea et al., 1998; Krag et al., 1998; Veronesi et al., 1999). For those with a negative SLN biopsy by histopathologic exam, the risk of “missed” axillary disease is extremely low (Giuliano et al., 2000; Turner et al., 1997). Therefore, SLN biopsy alone, without complete axillary lymph node dissection (ALND), has been adopted at many institutions as an accurate method of staging the axilla while avoiding much of the morbidity associated with a complete ALND. However, the standard of care for breast cancer patients with sentinel lymph node (SLN) metastases remains complete axillary lymph node dissection (ALND). Yet many question the need for completion ALND in every patient with detectable SLN metastases, particularly those in whom the perceived risk of additional disease is low (Chu et al., 1999; Kamath et al., 2001).
Proponents of completion ALND after a positive SLN biopsy argue that the further axillary clearance is critical to further management. The total number of involved nodes is important prognostic information, as an increasing number of positive nodes portends a worse survival (Cabanes et al., 1992; Moore et al., 1997; Carter et al., 1989). This is reflected in the new American Joint Committee on Cancer (AJCC) 6th edition staging system (2002), where the number of positive nodes defines N1, N2, and N3 disease and ultimately the stage to which the patient is assigned. In addition, proponents of complete ALND after positive SLN biopsy argue that the additional information can benefit patients by guiding decisions about adjuvant chemotherapy. For the approximately one-half of patients in whom there is residual nodal disease, it is also argued that complete ALND can influence survival via local-regional control of the axilla (Sosa et al., 1998; Hayward et al., 1987; Osteen et al., 1985), thereby eliminating a potential site of recurrent disease and, ultimately, a source for distant disease. A meta-analysis of randomized trials found a 5.4% survival benefit associated with ALND for clinically node-negative patients (Orr, 1999).
Opponents of complete ALND after positive SLN biopsy argue that the therapeutic benefit of complete ALND is minimal (Cady, 1997). Furthermore, approximately 50% of patients with positive SLNs are found to have no other nodal metastases (Giuliano et al., 1997; Albertini et al., 1996; Veronesi et al., 1997; O'Hea et al., 1998; Krag et al., 1998; Veronesi et al., 1999; Guiliano et al., 1994; Reynolds et al., 1999; Teng et al., 2000; Abdessalam et al., 2001; Rahusen et al., 2001). Therefore, many patients are possibly undergoing “unnecessary” ALND, with no additional therapeutic benefit or further staging information. It is also argued that because patients with SLN metastases will generally receive systemic therapy regardless of the presence of any additional nodal metastases, any residual disease does not influence choice of therapy and may itself be eradicated by the systemic therapy. In addition, radiation therapy after breast-conserving surgery may contribute to control of any additional nodal disease. It is this debate that physicians and their patients are faced with in the office setting when a positive SLN is discovered on final pathology.
Several groups have identified histopathologic variables of the primary tumor and its metastasis that can influence risk of having additional disease in the non-SLNs (Chu et al., 1999; Kamath et al., 2001; Reynolds et al., 1999; Teng et al., 2000; Abdessalam et al., 2001; Rahusen et al., 2001) (
Furthermore, it is difficult to estimate the risk of additional, non-SLN metastases for an individual patient using the published literature. First, the estimates of risk for any given characteristic vary considerably among studies.
What is needed is an improved method to predict additional nodal metastases in breast cancer patients with a positive sentinel node biopsy.
The invention provides methods, apparatus and nomograms to predict or determine the probability (likelihood) of additional nodal metastases in a breast cancer patient with a positive sentinel node biopsy. In one embodiment, the invention includes correlating the value or score from clinical and/or pathological data of the patient, for example, in a nomogram, to predict the likelihood of additional nodal metastasis. For instance, the methods, apparatus or nomograms may be employed after biopsy, e.g., after sentinel lymph node (SNL) biopsy, and before adjuvant therapy, and optionally prior to surgery for breast cancer, such as prior to completion axillary lymph node dissection, to predict the risk of additional nodal metastases in the patient.
As described herein, pathologic features of the primary tumor and sentinel lymph node metastases of 702 patients who underwent completion ALND were assessed with multivariable logistic regression to predict the presence of additional disease in the non-sentinel lymph nodes of these patients. A nomogram was created using pathologic size, tumor type and nuclear grade, lymphovascular invasion, multifocality, and estrogen receptor status of the primary tumor, as well as the method of detection of sentinel lymph node metastases, the number of positive sentinel lymph nodes, and the number of negative sentinel lymph nodes. The model was subsequently applied prospectively to 373 patients. The nomogram for the retrospective population was accurate and discriminating, with an area under the receiver operations curve (ROC) of 0.76. When applied to the prospective group, the model accurately predicted likelihood of non-sentinel lymph node disease (ROC of 0.77). Thus, a user friendly nomogram is provided which employs information commonly available to surgeons to easily and accurately calculate the likelihood of additional, non-sentinel lymph node metastases in an individual patient.
As also described herein, pathologic features of the primary tumor and SLN metastases of 33 patients who underwent completion ALND were assembled and presented to 17 breast cancer specialists. Their predictions for each woman were recorded and compared with results from the nomogram. The area under the ROC curve was computed for the nomogram, each clinician, and the clinicians as a whole. Subsequently, clinicians were presented with clinical information of 8 patients and asked if they would perform a completion ALND before and after being presented with the nomogram prediction. The predictive model achieved an area under the ROC curve of 0.72 when applied to the test data set of 33 patients. In comparison, the clinicians as a group were associated with an area under the ROC curve of 0.54 (P<0.01 vs. nomogram). When examined individually, one of the 17 clinicians outperformed the predictive model. Thus, the predictive model appeared to substantially outperform clinical experts.
Therefore, various factors from the primary tumor and/or sentinel lymph nodes, e.g., from a biopsy, in a breast cancer patient are employed to predict the likelihood of additional nodal metastases in that patient. In one embodiment, the prognosis is based on a computer derived analysis of data of the amount, level or other value (score) for one or more, e.g., two, three, four or more, of those factors. Data may be input manually or obtained automatically from an apparatus for measuring the amount, level or value of one or more factors.
Accordingly, the invention provides a method, apparatus, e.g., a computerized tool, and nomogram to predict the likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy, which is useful for counseling breast cancer patients. In one embodiment, the invention provides a method to determine a likelihood of additional nodal metastases in a breast cancer patient with a positive sentinel node biopsy. The method includes detecting or determining one or more factors, e.g., at least two, three, four, five, six, seven, eight or more factors, of the patient including but not limited to pathological size of the breast carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathologic evaluation of the sentinel lymph node, tumor type and nuclear grade, estrogen receptor status, presence or absence of lymphovascular invasion, and/or presence or absence of multifocality in the primary tumor, as well as expression and/or genomic data of the patient. The factors of the patient are correlated to the likelihood of additional nodal metastases in that patient.
Further provided is a method to determine the likelihood of additional nodal metastases in a breast cancer patient with a positive sentinel node biopsy, which includes inputting test information to a data input means. The information includes one or more factors, e.g., at least two, three, four, five, six, seven, eight or more factors, of the patient factors including but not limited to pathological size of the breast carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method of histopathologic evaluation of the sentinel lymph node, tumor type and nuclear grade, estrogen receptor status, presence or absence of lymphovascular invasion, and/or presence or absence of multifocality. A software is executed for analysis of the test information, and the test information analyzed so as to provide the likelihood of additional nodal metastases in the patient.
Also provided is a method for predicting the likelihood of additional nodal metastases in a breast cancer patient with a positive sentinel node biopsy. The method includes correlating one or more factors for the patient to a functional representation of one or more factors determined for each of a plurality of persons previously diagnosed with invasive breast cancer, having a positive sentinel lymph node biopsy and subjected to completion axillary lymph node dissection, so as to yield a value for total points for the patient. The factors for each of the plurality of persons is correlated with the likelihood of additional nodal metastases for each person in the plurality, wherein the one or more factors include but are not limited to the pathological size of the breast carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathogic evaluation of the sentinel lymph nodes, tumor type and nuclear grade, estrogen receptor status, presence or absence of lymphovascular invasion, and/or presence or absence of multifocality. The functional representation includes a scale for two or more of the factors, a total points scale, and a predictor scale. The scales for pathological size of the carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathogic evaluation of the sentinel lymph nodes, tumor type and nuclear grade, estrogen receptor status, lymphovascular invasion, and/or multifocality, each have values on the scales which can be correlated with values on the points scale, and the total points scale has values which may be correlated with values on the predictor scale. The value on the total points scale for the patient is correlated with a value on the predictor scale to predict the likelihood of additional nodal metastases in the patient.
The invention also provides an apparatus. In one embodiment, the apparatus includes a data input means, for input of test information for one or more factors from a breast cancer patient with a positive sentinel node biopsy, which factors include but are not limited to pathological size of the carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathogic evaluation of the sentinel lymph nodes, tumor type and nuclear grade, estrogen receptor status, the presence or absence of lymphovascular invasion, and/or the presence or absence of multifocality in the primary tumor, a processor, executing a software for analysis of each factor. The software analyzes the factors for the patient and provides the likelihood of additional nodal metastases in the patient.
In one embodiment, the invention provides an apparatus for predicting the likelihood of additional nodal metastases in a breast cancer patient with a positive sentinel node biopsy. The apparatus includes a correlation of one or more factors for each of a plurality of persons previously diagnosed with invasive breast cancer, having a positive sentinel lymph node biopsy and having completion axillary lymph node dissection, with the likelihood of additional nodal metastases for each person of the plurality of persons. The one or more factors include but are not limited to pathological size of the invasive carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathogic evaluation of the sentinel lymph nodes, nuclear grade, estrogen receptor status, presence or absence of lymphovascular invasion, and/or presence or absence of multifocality of the primary tumor. The apparatus also includes a means for comparing an identical set of factors determined from a patient having breast cancer and a positive sentinel lymph node biopsy to the correlation to predict the likelihood of additional nodal metastases in the patient.
Further provided is an apparatus for predicting the likelihood of additional nodal metastases in a breast cancer patient with a positive sentinel node biopsy. The apparatus includes a scale for one or more factors of the patient factor such as pathological size of the breast carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathogic evaluation of the sentinel lymph nodes, nuclear grade, estrogen receptor status, presence or absence of lymphovascular invasion, and/or presence or absence of multifocality in the primary tumor, a points scale, a total points scale and a predictor scale. The scales for pathological size of the breast carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathogic evaluation of the sentinel lymph nodes, nuclear grade, estrogen receptor status, presence or absence of lymphovascular invasion, and/or presence or absence of multifocality, each has values on the scales, and the scales for pathological size of the breast carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathogic evaluation of the sentinel lymph nodes, nuclear grade, estrogen receptor status, lymphovascular invasion, and/or multifocality, are disposed so that each of the values can be correlated with values on the points scale. The total points scale has values scale and is disposed on the solid support with respect to the predictor scale so that the values on the total points scale may be correlated with values on the predictor scale, such that the values on the points scale correlating with pathological size of the breast carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathogic evaluation of the sentinel lymph nodes, nuclear grade, estrogen receptor status, presence or absence of lymphovascular invasion, and/or presence or absence of multifocality of the patient can be added together to yield a total points value, and the total points value can be correlated with the predictor scale to predict the likelihood of additional nodal metastases in the patient.
Also provided is a system which includes a processor, an input device, an output device, a storage device, a database wherein the database includes data collected from a plurality of patients previously diagnosed with invasive breast cancer, having a positive sentinel lymph node biopsy and subjected to completion axillary lymph node dissection, and software operable on the processor to receive input from the input device. The input includes one or more factors for determining the likelihood of additional nodal metastases in a breast cancer patient with a positive sentinel node biopsy factors such as pathological size of the breast carcinoma, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, method for histopathologic evaluation of the sentinel lymph node, nuclear grade, estrogen receptor status, presence or absence of lymphovascular invasion, and/or presence or absence of multifocality in the primary tumor. The received input is correlated with the collected data from the plurality of patients to determine a likelihood of additional nodal metastases the patient.
The first row (POINTS) is the point assignment for each variable. Rows 2-9 represent the variables included in the model. For an individual patient, each variable is assigned a point value (uppermost scale, POINTS) based on the histopathologic characteristics. A vertical line is made between the appropriate variable value and the POINTS line. The assigned points for all eight variables are summed and the total is found in row 10 (TOTAL POINTS). Once the total is located, a vertical line is made between TOTAL POINTS and the final row, Row 11 (Predicted Probability of +LN).
The present invention provides methods, apparatus and nomograms to predict the likelihood of additional nodal metastases in a breast cancer patient with a positive sentinel node biopsy using factors available post-sentinel lymph node biopsy to aid patients considering completion axillary lymph node dissection. In one embodiment, a nomogram predicts the likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy using patient specific factors to assist the physician and patient in deciding whether or not the patient may benefit from completion axillary lymph node dissection.
One embodiment of the invention is directed to a post-sentinel lymph node biopsy method for predicting the likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy. The method includes correlating one or more factors, e.g., at least two, three, four, five, six, seven, eight or more factors, determined for each of a plurality of persons previously diagnosed with breast cancer, having a positive sentinel lymph node biopsy and subjected to completion axillary lymph node dissection, with the likelihood of additional nodal metastases for each person of the plurality to generate a functional representation of the correlation. In alternative embodiments, one or more subgroups of any one or more of the following factors may be excluded. The factors include but are not limited to pathologic size of the primary tumor, number of positive sentinel lymph nodes, number of negative sentinel lymph nodes, the method of detection (e.g., frozen, routine, serial HE or IHC), lymphovascular invasion, and/or multifocality, wherein the plurality of persons includes females having undergone sentinel lymph nodes biopsy, having a positive sentinel lymph nodes and completion axillary lymph node dissection. An identical factor or set of factors determined from a breast cancer patient having undergone sentinel lymph node biopsy is employed with the functional representation to predict the likelihood of additional nodal metastases in the breast cancer patient.
In one embodiment, the correlating includes accessing a memory storing the selected set of factors. In another embodiment, the correlating includes generating the functional representation and displaying the functional representation on a display. In one embodiment, the displaying includes transmitting the functional representation from a source. In one embodiment, the correlating is executed by a processor or a virtual computer program. In another embodiment, the correlating includes determining the selected set of factors. In one embodiment, determining includes accessing a memory storing the set of factors from the patient. In another embodiment, the method further comprises transmitting the quantitative probability of additional nodal metastases. In yet another embodiment, the method further comprises displaying the functional representation on a display. In yet another embodiment, the method further comprises inputting the identical set of factors for the patient within an input device. In another embodiment, the method further comprises storing any of the set of factors to a memory or to a database.
In one embodiment, the nomogram is generated with a Cox proportional hazards regression model (Cox, 1972, the disclosure of which is specifically incorporated by reference herein). This method can predict survival-type outcomes using multiple predictor variables. The Cox proportional hazards regression method estimates the probability of reaching a certain end point, such as disease recurrence, over time. In another embodiment, the nomogram may be generated with a neural network model (Rumelhart et al., 1986, the disclosure of which is specifically incorporated by reference herein). This is a non-linear, feed-forward system of layered neurons which backpropagate prediction errors. For instance, an artificial neural network (Dreiseitl et al., 2002, the disclosure of which is specifically incorporated by reference herein) or a Bayesian neural network (Barlow et al., 2001; Hauben et al., 2003, the disclosures of which are specifically incorporated by reference herein) may be employed. In another embodiment, the nomogram may be generated with a recursive partitioning model (Breiman et al., 1984, the disclosure of which is specifically incorporated by reference herein). In yet another embodiment, the nomogram is generated with support vector machine technology (Cristianni et al., 2000; Hastie, 2001, the disclosures of which are specifically incorporated by reference herein). In yet another embodiment, classification and regression trees (CART) can be used (Province et al., 2001; Begg, 1986, the disclosures of which are specifically incorporated by reference herein). Other models known to those skilled in the art may alternatively be used. In one embodiment, the invention includes the use of software that implements Cox regression models or support vector machines to predict the likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy.
The nomogram may comprise an apparatus for predicting the likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy. The apparatus comprises a correlation of one or more factors determined for each of a plurality of persons previously diagnosed with breast cancer, having a sentinel lymph node biopsy and having completion axillary lymph node dissection with the incidence of additional nodal metastases for each person of the plurality of persons. The factors include but are not limited to the pathologic size of the primary tumor, the number of positive sentinel lymph node, the number of negative sentinel lymph node, the method of detection (e.g., frozen, routine, serial HE or IHC), nuclear grade, estrogen receptor status, the presence or absence of lymphovascular invasion, and/or the presence or absence of multifocality. The apparatus includes a means for matching an identical set of factors determined from the patient having breast cancer to the correlation to predict the likelihood of additional nodal metastases in the patient with a positive sentinel node biopsy.
The nomogram or functional representation may assume any form, such as a computer program, e.g., in a hand-held device, world-wide-web page, e.g., written in FLASH, or a card, such as a laminated card. Any other suitable representation, picture, depiction or exemplification may be used. The nomogram may comprise a graphic representation and/or may be stored in a database or memory, e.g., a random access memory, read-only memory, disk, virtual memory or processor.
The apparatus comprising a nomogram may further comprise a storage mechanism, wherein the storage mechanism stores the nomogram; an input device that inputs the identical set of factors determined from a patient into the apparatus; and a display mechanism, wherein the display mechanism displays the quantitative likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy. The storage mechanism may be random access memory, read-only memory, a disk, virtual memory, a database, and a processor. The input device may be a keypad, a keyboard, stored data, a touch screen, a voice activated system, a downloadable program, downloadable data, a digital interface, a hand-held device, or an infra-red signal device. The display mechanism may be a computer monitor, a cathode ray tub (CRT), a digital screen, a light-emitting diode (LED), a liquid crystal display (LCD), an X-ray, a compressed digitized image, a video image, or a hand-held device. The apparatus may further comprise a display that displays the quantitative likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy, e.g., the display is separated from the processor such that the display receives the quantitative likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy. The apparatus may further comprise a database, wherein the database stores the correlation of factors and is accessible by the processor. The apparatus may further comprise an input device that inputs the identical set of factors determined from the patient with breast cancer into the apparatus. The input device stores the identical set of factors in a storage mechanism that is accessible by the processor. The apparatus may further comprise a transmission medium for transmitting the selected set of factors. The transmission medium is coupled to the processor and the correlation of factors. The apparatus may further comprise a transmission medium for transmitting the identical set of factors determined from the patient with breast cancer, preferably the transmission medium is coupled to the processor and the correlation of factors. The processor may be a multi-purpose or a dedicated processor. The processor includes an object oriented program having libraries, said libraries storing said correlation of factors.
In one embodiment, the nomogram comprises a graphic representation of a likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy. The nomogram comprises a substrate or solid support, and a set of indicia on the substrate or solid support, the indicia including one or more of a line for the pathological size, number of positive SLN, number of negative SLN, method of detection, lymphovascular invasion, and multifocality, a total points line and a predictor line, wherein the line for the pathological size, number of positive SLN, number of negative SLN, method of detection (frozen, routine, serial HE or IHC), lymphovascular invasion, and multifocality, each have values on a scale which can be correlated with values on a scale on the points line. The total points line has values on a scale which may be correlated with values on a scale on the predictor line, such that the value of each of the points correlating with the indicia can be added together to yield a total points value, and the total points value correlated with the predictor line to predict the likelihood of additional nodal metastases in breast cancer patients with a positive sentinel node biopsy. The solid support may assume any appropriate form such as, for example, a laminated card. Any other suitable representation, picture, depiction or exemplification may be used.
In addition to assisting the patient and physician in selecting an appropriate course of therapy, the nomograms of the present invention are also useful in clinical trials to identify patients appropriate for a trial, to quantify the expected benefit relative to baseline risk, to verify the effectiveness of randomization, to reduce the sample size requirements, and to facilitate comparisons across studies.
A block diagram of a computer system that executes programming for predicting a prognosis probability is shown in
Computer-readable instructions stored on a computer-readable medium are executable by the processing unit 1002 of the computer 1010. A hard drive, CD-ROM, and RAM are some examples of articles including a computer-readable medium. The computer-readable instructions allow computer system 1000 to provide generic access controls in a computer network system having multiple users and servers, wherein communication between the computers includes utilizing TCP/IP, COM, DCOM, XML, Simple Object Access Protocol (SOAP), and Web Services Description Language (WSDL), and other related connection communication protocols and technologies that will be readily apparent to one of skill in the relevant art.
In some embodiments, one or more of the servers 1102 hold a prediction program 1104, which is available for download to the other servers 1102 and workstation clients 1106 connected 1124 to the network 1122.
In some other embodiments, prediction program 1104 is executable on a server 1102 wherein the prediction program executes in response to stimulation received from a client 1104 using a Hyper-Text Transfer Protocol (HTTP or HTTPS). In one such embodiment, prediction program 1104 accepts input from a client, executes, and outputs a prognosis prediction in a markup language such as Hyper-Text Markup Language (HTML) or extensible Markup Language (XML).
In some embodiments, system 1100 may be implemented with servers 1102 utilizing one of many available operating systems. Servers 1102 may also include, for example, machine variants such as personal computers, handheld personal digital assistants, RISC processor computers, MIP single and multiprocessor class computers, and other personal, workgroup, and enterprise class servers. Further, servers 1102 may also be implemented with relational database management systems 1103 and application servers. Other servers 1102 may be file servers.
Client workstations 1102 within embodiments of system 1100, may include personal computers, computer terminals, handheld devices, and multifunction mobile phones. Client workstations 1102 include software thereon for performing operations in accordance with stimulation received from a user and signals received from other computing devices on the network 1122. Further, a client workstation 1102 may include a web browser for displaying web pages.
The network 1122 within some embodiments of a system 1100 may include a Local Area Network (LAN), Wide Area Network (WAN), or other similar network 1145 connected 1124 network 1122. Network 1122 may itself be a LAN, WAN, the Internet, or other large scale regional, national, or global network or a combination of several types of networks. Some embodiments of system 1100 include a LAN, WAN, or other similar network 1145 that utilizes one or more servers 1152 and clients 1155 behind a firewall 1160 within the LAN, WAN, or other similar network 1145.
The invention will be further described by the following non-limiting examples.
In an attempt to achieve a more precise prediction for the individual patient than is readily available by using published estimates of risk, a multivariable logistic-regression analysis of a large data set was used to model the association between selected variables and the likelihood of metastases in non-SLN in patients with a positive SLN biopsy. The pathologic size of the primary tumor, the method of detection of the SLN metastasis, as well as several other variables that are readily available and likely related to risk of additional nodal disease, were examined. 702 cases from a large prospective sentinel lymph node database were employed to develop the model, and a nomogram was developed to predict the likelihood of finding additional positive nodes at completion ALND. The model was then tested by prospectively applying it to an additional study group comprising 373 patients. This tool allows greater individualization of a patient's risk estimate by simultaneously taking into account several pertinent characteristics specific to the patient. With a more precise and individualized estimate, both physician and patient are better able to weigh the pros and cons of further axillary dissection.
Materials and Methods
4,790 consecutive cases of SLN biopsy at Memorial Sloan-Kettering Cancer Center (MSKCC) were entered prospectively into the MSKCC Breast Cancer Sentinel Lymph Node Database. The study population was the subset of 1,075 that fulfilled the following criteria: primary invasive breast carcinoma with clinically negative axilla and no prior systemic treatment, successful SLN biopsy in which metastatic disease was identified, and completion ALND with at least 10 nodes examined. There were a total of 140 cases that were excluded because a completion ALND was not performed. Patients meeting selection criteria included the overall study population, which was then divided into two groups: a retrospective group who had undergone SLN biopsy over about 4½ years, and a prospective group undergoing SLN biopsy for the subsequent 1½ years. This project was reviewed and approved by the MSKCC Institutional Review Board.
The technique for SLN biopsy includes the use of both blue dye and radioisotope as described in Cody (2001).
SLN Histopathologic Evaluation
Whenever possible, the SLN was bisected and sectioned at 2- to 3-mm intervals. The nodal tissue was quick frozen in liquid nitrogen and a single, 5-μm-thick hematoxylin and eosin (H&E) stained section was examined intraoperatively (frozen-section analysis). If positive, a complete ALND was done immediately. Following the frozen section, the remaining frozen tissue was fixed in formalin and embedded in paraffin. Another 5-μm-thick H&E-stained section was evaluated as a “frozen section control” (routine histopathology). If this section showed evidence of metastatic disease, no further pathological workup of the SLN was performed. If the routine H&E section remained negative, enhanced pathologic analysis was performed in the following fashion: two pairs of H&E- and cytokeratin IHC-stained sections with a distance of 50 μm between the pairs were prepared from the paraffin block. At one level, the cytokeratin antibody CAM 5.2 (Becton Dickinson Immunocytometry Systems, San Jose, Calif.) was used, while the cytokeratin cocktail AE1:AE3 (Ventana Medical Systems, Inc., Tucson, Ariz.) was applied for the other level. Patients with SLN metastases not detected by frozen-section analysis generally underwent completion ALND at a later date. All additional nodes identified by completion ALND underwent routine H&E analysis of a single section of each node.
Data Analysis
Clinical data collected for each case from the database included age; pathologic size of the invasive carcinoma, defined in centimeters (cm); tumor type (ductal or lobular carcinoma); nuclear grade (I: slight or no variation in size and shape of nucleus; II: moderate variation in size and shape; III: marked variation in size and shape); presence of lymphovascular invasion (presence of a one or more tumor cells in a lymphatic or vascular structure); multifocality of primary tumor (foci of carcinoma separate from primary tumor); estrogen-receptor (ER-) status (negative, <10% of cells staining positive); method of detection of SLN metastases (frozen-section analysis, “Frozen”; routine histopathology, “Routine”; H&E stains of serial sections, “Serial HE”; immunohistochemistry, “IHC”); number of positive SLNs; and number of negative SLNs. Because lobular carcinomas are generally not assigned a nuclear grade, tumor type and nuclear grade were combined into the following four categories: ductal carcinoma, nuclear grade I; ductal carcinoma, nuclear grade II; ductal carcinoma, nuclear grade III; lobular carcinoma.
To allow use of the model by groups that do not routinely perform frozen-section analysis, a second model was developed with only three levels for the method of detection variable: routine histopathology, serial sectioning, and IHC. In this model, a node in which metastatic disease was detected by either frozen-section analysis or routine histopathology was categorized as “routine”. Data on additional variables such as progesterone-receptor status, histologic grade, and AJCC T stage were also collected; however, because these variables are highly correlated with ER-status, nuclear grade, and pathologic size, respectively, it was felt they would not be of substantial benefit to the model. HER-2/neu amplification data were also collected, but were not included because they were incomplete and variable owing to evolving methods of assessment during the years of the study.
A nomogram was developed based on the patients in the retrospective group, and then validated with the patients in the prospective group. In the retrospective population (n=702), multivariable logistic regression was used to analyze the association of each variable with the likelihood of non-SLN metastases, and a nomogram was created with all variables. This model was used in the prospective group (n=373) to predict each individual patient's probability of having positive non-SLNs. The discrimination of the model was measured by using the area under the receiver operating characteristic (ROC) curve. The calibration of the model was assessed graphically. Women were grouped into deciles based on their nomogram predictions. For each decile, the mean nomogram-predicted probability was compared with the proportion of women who actually had positive non-SLNs (actual probability). All analyses were performed using S-Plus software Version 2000 Professional Edition with the Design Library (Mathsoft Data Analysis Products Division, Seattle, Wash.) (Harrell, 2001).
Results
Descriptive characteristics of the study population are listed in Table 1.
SLN, sentinel lymph node.
Table 2 shows the incidence of additional, non-SLN metastases for retrospective, prospective, and total patient populations by primary and SLN pathologic characteristics.
SLN, sentinel lymph node.
On multivariable logistic-regression analysis, pathologic size, lymphovascular invasion, method of detection, number of positive SLNs, and number of negative SLNs, were each associated with the likelihood of additional, non-SLN metastases (P<0.05 for each). Multifocality was of borderline significance, and neither tumor type and nuclear grade nor ER status had a statistically significant association with the likelihood of non-SLN metastases (Tables 3 and 4).
HE, hematoxylin and eosin;
IHC, immunohistochemistry,
SLN, sentinel lymph node.
HE, hematoxylin and eosin;
IHC, immunohistochemistry,
SLN, sentinel lymph node.
Age was not included in the final nomograms because its effect was too small to be seen on the nomograms.
A nomogram based on this model and developed in the retrospective population (n=702) is shown in
Using the Nomogram
Each version of the nomogram consists of 11 rows. The first row (POINTS) is the point assignment for each variable. Rows 2-9 represent the variables included in the model. For an individual patient, each variable is assigned a point value (uppermost scale “POINTS”) based on the histopathologic characteristics. To determine the point assignment, a vertical line is made between the appropriate variable value and the POINTS line. For example, a pathologic size of 1 cm (PATHSIZE, 2) confers about 10 points.
The assigned points for all eight variables are summed, and the total is found in row 10 (TOTAL POINTS). Once the total is located in row 10, a vertical line is made between it and the corresponding value in the final row, Row 11 (Predicted Probability of +LN). The version of the nomogram in
In addition to the graphical nomograms, to facilitate ease of use in the clinical setting, a personal digital assistant (PDA)-compatible application for use on hand-held Palm™-type devices is provided at www.mskcc.org/nomograms.
Discussion
With the adoption of SLN biopsy, a new clinical conundrum has become commonplace: should a completion ALND be done for a patient with a positive SLN biopsy? This question is particularly difficult in patients with micrometastatic disease, disease which was undetectable in the era prior to SLN biopsy. Other investigators have attempted to address this question, and have identified risk factors for the presence of additional, non-SLN disease, but all such attempts are limited by the practical difficulty of simultaneously including several variables in the risk estimate. Here, simultaneously using several variables in a large population, nomograms were developed to predict the likelihood of additional nodal metastases after a positive SLN biopsy. The nomograms were prospectively tested and shown to perform well in the prospective population.
The nomograms utilize available clinical information, and allow quick calculation. This approach may allow identification of extremely low-risk individuals in whom the risks associated with completion ALND are judged to outweigh the benefits. On the other hand, the nomogram may allow identification of women at sufficient risk of additional nodal disease that they and their surgeon elect to proceed with completion ALND even though clinical “guesstimates” would suggest that they are at “low risk.”
The nomograms provide risk estimates that are judged on an individual basis. A woman with a 1.8-cm, ER-positive, high-nuclear-grade ductal carcinoma with no LVI, who has a single IHC-positive SLN might be considered to be “low risk.” The nomogram suggests that she has a 12% risk of having non-SLN metastases. Should she undergo completion ALND? Given this scenario, some will judge that a 12% risk of additional, non-SLN metastases justifies further ALND, others will not. The nomogram itself makes no actual treatment recommendations.
The nodes retrieved at completion ALND were examined by routine pathologic analysis only. Other investigators (Turner et al., 2000; Chu et al., 1999) have shown that if non-SLNs are examined with serial sectioning and IHC, a higher proportion of patients with additional, non-SLN disease at completion ALND are identified.
Moreover, the clinical relevance of resecting additional nodal disease (even that detected by routine analysis) remains unknown. While some argue that surgical removal of subclinical nodal disease is associated with a small, but non-zero, survival benefit, others argue that current adjuvant systemic therapy and radiation therapy would likely treat the majority of patients adequately. This study does not address this issue, but rather provides accurate and individualized estimates of the likelihood there will be additional disease at completion ALND. The American College of Surgeons Oncology Group Protocol Z0011 (ACOSOG Z0011), currently under way and randomizing women with a positive SLN to ALND or no, is designed to address this question directly.
In addition, the prognostic significance of micrometastatic nodal disease is a subject of debate. In a 1997 review of the published literature, Dowlatshahi et al. (1997) concluded that all but one of the large (N≧147) and long-term (≧6 yrs) studies demonstrated a statistically significant decrement in survival associated with micrometastatic disease. Tan et al. (2002) recently re-examined all axillary nodes from 373 patients treated in the 1970s who were deemed to be node-negative by routine histopathologic analysis. Nodes were examined by serial sectioning and IHC, and the presence of any detectable micrometastatic disease was associated with worse disease-free and overall survival.
The AJCC Cancer Staging Manual, 6th edition, now includes size of metastasis as an important determination of stage (and, therefore, of prognosis). However, it can be difficult to assign a size to many cases because of the difference in pattern of distribution of malignant cells within the node. For example, some nodes may have scattered single cells or multiple small clusters of cells. How should these be measured? Ideally, an accurate estimate of volume could be assigned to each SLN metastasis. However, this is extremely time-consuming and somewhat impractical.
Nevertheless, it is clear that IHC is more sensitive than H&E in detecting micrometastases, that routine H&E analysis is more sensitive than frozen-section analysis, and that there is a correlation between method of detection and volume of disease. Kamath et al. (2001) and Rahusen et al. (2001) have demonstrated quantitatively that method of detection is correlated with measured size of the SLN metastasis. Therefore, in order to have a consistent, practical, and reproducible methodology of estimation, the method of detection of the nodal metastasis was used. This provides a general estimate of the amount of nodal disease, and allows grouping into four distinct groups.
Some of the patients in the present study, especially those with a perceived low risk of additional, non-SLN metastases, did not have a completion ALND and therefore were not included in the model. However, as demonstrated in the histograms (
The models described herein represent a significant improvement over estimates based on one or two variables in smaller populations. A large, prospective database was used to develop the models, and their validity proven by testing them prospectively on a subsequent population. The calibration errors of the models are small (see
With the important clinical question of whether to perform a completion ALND in a patient with a positive SLN biopsy arising more and more frequently, the present nomograms provide an easy-to-use tool with which to simultaneously incorporate several important variables into the estimate of risk of additional, non-SLN metastases. These nomograms provide a risk estimate that can help in weighing the pros and cons of completion ALND for an individual patient with SLN metastases.
Methods
Construction of the nomogram is described in Example I. In brief, 702 cases of primary breast cancer in which the SLN was positive for metastasis were identified from a prospectively collected SLN database. Using primary tumor and SLN metastasis characteristics, a multivariate model was created to predict the likelihood of additional, non-SLN metastases being found at completion ALND. The model was subsequently applied prospectively to an additional 373 patients (validation population), and found to accurately predict the likelihood of residual disease (area under the receiver operating characteristic curve=0.77).
For experiment I, 33 women were selected at random from the validation population used to confirm the original nomogram. The characteristics of these women were supplied to 17 participating clinicians for their prediction (Table 5). Clinicians were asked, for each patient, “If 100 women with these characteristics were to have a positive sentinel node and then receive a full axillary dissection, how many of them would you expect to have one or more positive non-sentinel lymph nodes?” Clinicians included specialists in breast cancer who attended a weekly multidisciplinary breast conference. Individual specialties included: surgeons, medical oncologists, radiation oncologists, radiologists, and pathologists. These clinicians were unfamiliar with the nomogram and had not yet incorporated it into their clinical practice at the time they participated in this experiment.
Subsequently, in experiment II, clinicians were presented with tumor and patient characteristics from 8 patients in the validation set. Clinicians were specifically asked, “Would you perform a completion axillary dissection on this patient?” After they answered, they were then presented the results of the nomogram prediction for residual axillary disease in that patient. Clinicians were then asked again, “Would you perform a completion axillary dissection?”
Twenty-four clinicians participated in experiment II. This study was conducted during a multidisciplinary breast cancer conference. The room was equipped with audience participation software and each participant was given a handheld device and responses were recorded for each device. Unfortunately, not all participants responded to all questions. There were 187 responses to 8 questions, which resulted in a 97.4% response rate.
For experiment I, the clinician and nomogram predictions were evaluated by calculating the area under the receiver operating characteristic curve (ROC). The ROC curve is a visual representation of the tradeoff between sensitivity and specificity of a diagnostic test. This curve describes the inherent predictive ability of the test. Each point along the curve corresponds to the sensitivity and specificity for that test threshold. In the present study, the curve describes the sensitivity and specificity of the nomogram at different levels of likelihood of residual disease.
The area under the ROC curve (AUC) is a value measurement that allows the comparison of the predictive ability of two tests. If the AUC is 0.5, then the curve is approximately a straight line, and the test is no better than a flip of the coin in predicting the desired outcome. However, if the AUC is 1.0, the test is perfect, and correctly identifies all the true positives and true negatives.
To compare the accuracy of nomogram and clinician predictions of lymph node positivity, an ROC curve approach was employed. Specifically, the parameters of the ROC curves were estimated using a latent-variable binormal model (Temple et al., 2002). A random effects term was added to account for the fact that each patient was evaluated several times by different physicians. Model estimates were obtained using restricted maximum likelihood with SAS PROC MIXED (SAS/STAT User Manual Version 9, 2003) and the accuracy of nomogram and clinician predictions were compared using a likelihood ratio test.
For experiment II, a mixed model was used, this time to evaluate whether a statistically significant shift in clinician judgment had occurred upon viewing the nomogram predictions.
Results
In experiment I, 17 clinicians made predictions for 33 patients. The nomogram predicted more accurately (AUC=0.72) than the clinicians (AUC=0.54) as a group (P<0.01). One clinician outperformed the nomogram while the remaining 16 made predictions that were inferior to those made by the nomogram.
In experiment II, 24 clinicians responded to 8 questions, which resulted in 187 responses, 97% response rate. Clinicians rarely changed their surgical decision after being presented with the nomogram prediction of non-SLN metastases (Table 6). Ninety percent of responses represented no change in surgical plan (168/187). Among those recommendations for not proceeding with completion ALND, half (7/14) were changed after being presented with the nomogram prediction. Among recommendations for completion ALND, only 7% (12/173) were changed to “no ALND” after presentation of the nomogram prediction.
Discussion
Outside of a clinical trial, completion ALND after a positive sentinel lymph node biopsy is recommended. However, for patients in whom the perceived risk of residual disease is low, some patients and clinicians feel that the benefit of axillary dissection is outweighed by its risks, and choose not to have a completion ALND. For such patients, the nomogram was developed to provide an accurate risk estimate that can help in weighing the pros and cons of completion ALND.
Uncertainties in medical decision-making are plentiful, including inaccuracies in diagnosis, uncertainties in the natural progression of disease, and variations with regards to the effect of treatment in an individual patient. In discussing the risks and benefits of performing a completion ALND after a positive SLN biopsy, the physician must process abundant data to arrive at their best prediction of the likelihood of finding residual disease. As demonstrated herein, when presented with similar input data to answer this question, a statistical model (nomogram) predicted more accurately than clinical experts the status of the axilla after a positive SLN biopsy.
This study supports a previous finding that nomogram models outperform human experts (Ross et al., 2001). Humans are filled with inherent biases that make predicting outcomes difficult. Clinicians are plagued by recall bias, remembering the unique patient rather than the routine. Control bias occurs when outcomes are predicted that one wants to come true. Practically, clinicians utilize simple rules to stratify patients rather than a continuous regression analysis (Ross et al., 2001). For example, a clinician heuristic might be that tumors over 5 cm in size are “high risk” for nodal metastases rather than utilize size as a continuous variable. Therefore, it is not surprising that the nomogram did perform better than clinical experts. The nomogram is a predictive instrument that can accurately weigh multiple individual variables simultaneously and without bias.
Does this mean that the nomogram should replace the clinical expert in making the decision about completion ALND after a positive SLN biopsy? Of course not, but the nomogram could be added as a tool to the decision making process in providing the clinician with a numerical estimate to help both the clinician and patient weigh the pros and cons in making this decision. There is no inherent ability of a nomogram to perform risk/benefit analyses and therefore it can not replace clinical judgment.
Outside of a clinical trial, completion ALND is recommended after a positive sentinel lymph node biopsy. However, some clinicians and patients may feel that the risk of residual disease is low and that the morbidity of ALND high and therefore may forgo dissection based on this information. The nomogram provides such patients an accurate estimate of the likelihood of residual disease and thereby allow an informed decision.
In experiment II indicates that the nomogram was unable to change physicians' behavior. It is possible that the clinicians were unfamiliar with the nomogram as a prediction tool and that with increased use would be more comfortable relying on the results to be part of the decision making process. Alternatively, reported clinical decisions may not be reliable outside of the clinic and therefore only with a patient present could a true estimate of change in behavior be made. However, the most likely explanation is that clinicians believe that the standard of care is completion ALND after a positive SLNB, and only rarely consider not recommending a completion ALND (14/187=7%). It is informative that of these 14 responses indicating a preference for not doing completion ALND, fully half (7/14=50%) changed their minds to recommend completion ALND after hearing the nomogram estimate of likelihood of residual disease. Conversely, of the 173 recommendations for ALND, only 12 (12/173=7%) were changed to recommend no ALND after hearing the nomogram estimates.
Interestingly, in two scenarios, the decision whether or not to dissect the axilla was different, even though the risk of residual disease was the same. A post-menopausal woman with a 9% risk of residual disease was presented and two clinicians changed their behavior to not perform completion ALND; presumably the nomogram prediction was lower than what the clinician had anticipated. Conversely, a 38-year-old woman with an 8% risk of residual disease was presented and four clinicians changed their response to dissect the axilla—presumably the nomogram prediction was higher than they had predicted. Clearly, age plays a role in making the decision about returning to the operating room. However, in constructing the nomogram, age had no predictive role in determining likelihood of non-SLN metastases. A patient's age is seemingly associated with the amount of risk clinicians are willing to assume.
The question of estimating the risk of non-SLN metastases in the axilla after a positive SLN biopsy is an important one that clinicians are facing more frequently. Accurate estimates of this likelihood may improve risk stratification in future clinical trials of the utility of completion ALND. Further, accurate estimates of risk are essential for an informed discussion with patients regarding the pros and cons of completion ALND in the setting of a positive SLN biopsy. Nomogram predictions appear to be substantially more accurate than clinical predictions, and therefore clinicians can improve their predictive ability by using the nomogram to predict the likelihood of additional non-SLN metastases in a woman with a positive SLN biopsy who is considering not pursuing completion ALND.
All publications, patents and patent applications are incorporated herein by reference. While in the foregoing specification, this invention has been described in relation to certain preferred embodiments thereof, and many details have been set forth for purposes of illustration, it will be apparent to those skilled in the art that the invention is susceptible to additional embodiments and that certain of the details herein may be varied considerably without departing from the basic principles of the invention.
This application claims the benefit of the filing date of U.S. application Ser. No. 60/525,325, filed Nov. 26, 2003, under 35 U.S.C. § 119(e), the disclosure of which is incorporated by reference herein.
| Number | Date | Country | |
|---|---|---|---|
| 60525325 | Nov 2003 | US |
